Changelog (2026)¶

This document records the release notes for each update of Guance.

DataKit

SDK

Web Mini Programs Android iOS React Native

August 26, 2026¶

Feature Updates¶

Obsy AI¶

AI API Call Billing: AI API calls are now included in the Task Invocation Count metric and billed by invocation, with usage recorded separately by sub-project.
Agent Identifiers in Alert Policies: Alert policies now display a dedicated identifier for the Agent type, making policies associated with Agents easier to recognize.

Obsy Agent Teams¶

Improved Built-in Business Skills: Optimized built-in OWL Skills around business scenarios, analysis logic, and output standards. By reducing interference from low-level tool details, Agents can invoke tools more reliably and accurately.
Improved Task State and Response Recovery: Standardized the task lifecycle and status definitions across queuing, execution, approval waiting, and completion. Task status, elapsed time, and progress information now remain consistent during real-time updates, page refreshes, and reconnections.
Improved Chinese and English Output: Enhanced Agent language detection and matching so that Chinese and English responses are more accurate and consistent, improving readability and interaction quality.

Incident Center¶

Incident War Rooms: Users can create a Feishu group chat or Tencent Meeting directly from incident details to quickly bring responders together.
- Feishu war rooms synchronize incident status and assignees, and allow members to claim or close incidents from cards in the group chat.
- Tencent Meeting war rooms support quickly creating and joining meetings.

Agent Monitoring¶

LangChain Observability Data Compatibility: Agent Monitoring can now parse input, output, Token, and other fields from LangChain OpenTelemetry data, improving data presentation and analysis.
Agent / LLM Evaluation: Added Prompt Eval, Schema Eval, online evaluation, batch evaluation, and evaluation result analysis.

Status Page¶

Improved Service Status Display: Enhanced service status and historical incident displays, with support for viewing status history from the past 90 days and the current SLA.

Application Performance Monitoring¶

Service Topology Grouping: Services can now be grouped and visually framed by Kubernetes cluster, environment, and the project tag while preserving calls across groups.
Project-Level Separation for Services with the Same Name: Services with the same name can now be separated by project, preventing data from different projects from being merged.
Improved Detection of Uninstrumented External HTTP Services: Expanded the lookup logic for Root Spans and parent Spans. If no parent can be found but a valid destination address is available, a named virtual Gateway is generated in the service call map.
- Virtual Gateways do not currently display metrics and are not yet supported in Waterfall or Flame Graph views.

Logs¶

Improved Clone Index Display and Queries: Query entry points now display standard indexes and their clone indexes in a tree structure, and users can select a clone index directly for queries.
Enhanced Field Value Filters: The Log Explorer can now display and filter null values separately from empty strings (""), improving query accuracy for special field values.

Real User Monitoring¶

Improved RUM Setup Flow: When creating an application, users first provide the application type, name, application ID, and Client Token before proceeding to the setup page, making application configuration and integration steps clearer.

Dashboards¶

Improved Dashboard Visuals: Updated the official chart color palette and introduced bordered group layouts to present charts in the same group together.
Smart Group Layout: Added one-click automatic arrangement of charts within a group, improving dashboard layout efficiency.

Infrastructure¶

Expanded MongoDB Database Analysis: MongoDB databases now include Metrics and Query Analysis tabs, extending database monitoring and query analysis capabilities.

Cloud Account Management¶

GCP and Volcengine Cloud Accounts: Google Cloud and Volcengine accounts can now be connected, extending centralized management across multi-cloud environments.

Ticket Management¶

Sensitive Information Warnings: When users create a ticket or submit a comment, the system identifies potentially sensitive content and prompts them to confirm submission or return to make changes.
Enhanced Permissions and Auditing: Strengthened permission controls for deleting tickets and added audit records for ticket creation and deletion.

Workspace¶

Improved Lock Status Messaging: Optimized page messages for the 14-day workspace lock countdown and fully locked state, clearly explaining the workspace status, usage restrictions, and available actions.
Improved Lock Notification Emails: Updated the Chinese and English email content for data ingestion suspension, full workspace lock, and service restoration.

Other¶

Console API Rate Limiting: Added general rate-limiting capabilities for console APIs. Access frequency limits and shared quotas can now be configured for selected endpoints across OpenAPI, Inner API, External API, and other entry points.

Bug Fixes¶

Fixed an issue where log monitors queried historical aggregate values outside the configured time range, causing false alerts;
Fixed an issue where log queries still performed tokenized searches when keywords were enclosed in double quotes;
Fixed an issue where session replay data was available but could not be played;
Fixed an error that occurred when opening shared log snapshots;
Fixed an issue where some chart names were not displayed when a dashboard had no data;
Fixed an issue where the total Token count was not displayed for some Traces in LLM observability scenarios;
Fixed an issue where an application with the same app_id and client_token as an existing application could be created via API across workspaces within the same site;
Fixed an issue where configured view variables were not correctly applied when scheduled reports delivered views;
Fixed incorrect P90 calculations when dashboards queried frontend data;
Fixed missing data in some charts;
Fixed external data source connection failures.

August 19, 2026¶

Feature Updates¶

Obsy AI¶

Copilot Analysis for Infrastructure: Infrastructure page context can now be passed to Copilot for intelligent analysis. A new AI Analysis shortcut helps users quickly identify issues and obtain analysis results.

Obsy Agent Teams¶

Self-Observability Changes Take Effect Immediately: Changes to self-observability options now take effect immediately, without restarting or rerunning the Agent Team, making debugging and daily use more efficient.

Logs¶

Log Index Cloning: After a log is written to the first matching standard index, it can be asynchronously copied to a dedicated clone index based on additional conditions. This reduces the data scope of frequent queries and improves search efficiency.
- When creating or editing a log index, users can configure the clone index name, retention period, and matching conditions.
- A clone index retains all fields and tags from the original log. Each standard index can have one clone index, and cloning failures do not affect writes to the primary index.
- Clone rules apply only to logs written after the rules take effect and do not backfill historical data.
Expanded Search Highlighting: Search highlighting is no longer limited to the message field. All content matching the search criteria is highlighted, making keywords easier to locate.
Improved Analysis View Switching: When a BY grouping condition is configured in a Top List, switching to a List or Stacked List and then returning preserves the previous grouping condition and related analysis settings.
Improved Advanced Query Compatibility: If an advanced query contains a BY condition, the page ignores that condition to prevent conflicts with the page-level analysis configuration.

DQL¶

SHIFT Time-Offset Queries: Aggregate expressions can now use <aggregate> SHIFT <duration> to retrieve data from a historical window and align it with current results for period-over-period and trend comparisons. For example, count(*) SHIFT 1d returns the aggregate result from one day before the current window.

Data Forwarding¶

Agent Monitoring Data Source: Agent Monitoring can now be selected as a data source for Data Forwarding.
TiDB Cloud Lake Destination: Observability data can now be forwarded in Parquet format to a target table in TiDB Cloud Lake for long-term SQL queries and business data correlation.

Integrations¶

TiDB Cloud Lake Integration: TiDB Cloud Lake can now be connected directly as an external data source. Native SQL queries can be executed from existing query entry points such as Metric Analysis and Dashboards.

Infrastructure¶

Network Device Topology: Added a topology view that displays physical connections between network devices, providing a clearer view of device relationships.

Dashboards¶

Guaranteed Streaming Aggregation Acceleration: When enabled, eligible queries use streaming aggregation acceleration by default, improving response stability for fixed queries.
- The system automatically determines eligibility based on query characteristics. Queries that use complex regular expressions and certain other complex query patterns may not currently be covered. Eligibility is subject to the system's evaluation, and support will continue to expand.
Improved Top List Interactions: Optimized click interactions for Top List charts.

Generated Metrics¶

Query Acceleration List: Added a list that displays accelerated streaming aggregations and their source dashboards and charts, making acceleration configurations easier to trace and manage.

Agent Monitoring¶

Additional Agent Types: Agent Monitoring can now identify and display data from CodeBuddy, Spring AI Alibaba, and Cursor.

Application Performance Monitoring¶

Default Trace Details View: Preferences now include a default view option for Trace details. Users can open either the Waterfall or Flame Graph view by default, reducing repetitive switching during troubleshooting.

Work Area¶

Work Area: Multiple feature pages can now be retained and switched between as tabs, reducing context loss when navigating among Logs, Traces, Dashboards, and other modules.
- Users can add the current page to the Work Area or open a new page from the module selector or navigation menu.
- Up to eight work items can be open at the same time.

Other¶

API Key Notes: The API Key list now displays notes, making keys created for different purposes easier to identify and manage.
Free Plan Workspace Service Discontinuation Reminder: An upgrade notice is now displayed when users enter feature modules, reminding them that the Free Plan service will end on September 2, 2026, and to save important data in advance.

August 12, 2026¶

Feature Updates¶

Obsy AI¶

Copilot Model Selection: Users can select a model intelligence level based on task requirements, allowing Copilot to match an appropriate model while balancing analysis quality, response speed, and credit consumption.
Dashboard Analysis Optimization: Optimized analysis strategies for large time ranges, dashboards with many charts, and complex queries to reduce long waits and timeouts. Sampling coverage is now stated more clearly, improving the accuracy and readability of analysis results.

Obsy Agent Teams¶

Improved MCP Configuration: MCP can now be configured through a form or JSON, with enhanced tool list loading, viewing, and refresh capabilities for clearer visibility into the current MCP configuration.
Knowledge Base References: Knowledge Base has been added to the @ resource list in tasks, allowing Agents to reference knowledge content directly and use broader context during task execution.
Improved Resource Reference Parameters: Standardized resource entities and OWL tool parameters to reduce first-call failures when referencing resources such as Deployments and improve tool invocation accuracy.
Real-Time Event Message Forwarding: After receiving a workspace event, an Agent can forward the original event to an IM group with an analysis-in-progress notice, followed by the analysis result. This provides faster feedback and simplifies notification configuration.
Conversation Aliases for Messaging Channels: Aliases can now be assigned to conversations in WeChat messaging channels, making conversations easier to identify when the returned ID is unreadable.
Improved Credit Consumption Statistics: Optimized the query pipeline for credit consumption statistics to improve the loading performance and reliability of the statistics page.

Logs¶

Log Index Query History: Multi-index queries are now recorded in query history, with multiple index names displayed as a comma-separated list.
Index-Specific Display Items: Display fields can now be configured and saved independently for each log index. Changes to one index no longer affect other indexes.
- Full Row Mode now displays business fields by default instead of always displaying the message field.
- Log details dynamically display content based on the combination of business fields.
Enhanced Search Suggestions: The log search box now suggests common logical operators such as AND and OR, helping users write compound query conditions more quickly.
Enhanced CSV Export:
- The Grouped Table Chart export menu now distinguishes between Sampled CSV Export and Full CSV Export.
- Full export now provides options for 5,000 and 10,000 records.
- Exports of 5,000 and 10,000 records are not sampled and use the same export logic as the Log Explorer.

Application Performance Monitoring / Real User Monitoring¶

Improved Explorer Interactions: Added Quick Filter and Display Items switching, making it easier to adjust the fields displayed in lists.

Agent Monitoring¶

User Analysis Dimension: Related data can now be viewed and analyzed by User.
Skill Execution Percentage: Call Analysis on the Session details page now displays the execution percentage of each Skill, making Skill execution distribution easier to understand.

Administration¶

Metric Data Point Reporting Limit: Added a Data Point Reporting Limit setting under Manage > Workspace Settings > Risky Operations. Commercial workspaces billed by data points can set a daily metric data point limit. Once the limit is reached, the system stops ingesting metric data and automatically resumes at 00:00 the following day.
- Supports configuring the daily limit, in units of ten million data points, and email notification recipients.
- Sends an email notification when usage reaches 80% of the limit. At 100%, the system stops ingesting data and generates a system event.
- Enabling the setting requires an additional confirmation step, while disabling it saves the change directly.
- Applies only to workspaces billed by data points and is not displayed in Free Plan workspaces.
- Only workspace Owners, Administrators, and roles with the Manage > Risky Operations permission can configure this setting.

Data Forwarding¶

Apache Iceberg Storage Format for AWS S3 Archives: Data is written to external Iceberg tables in Parquet format through REST Catalog. The system automatically creates the Namespace and Table and partitions data by hour(date), enabling direct integration with data lake ecosystems such as Spark, Trino, and Flink.
- The Iceberg format currently supports write operations only. Archived data cannot yet be queried within the platform, but it can be read through an external query engine connected to the same Catalog.
- Added a Storage Format filter to the Data Forwarding rule list, allowing users to quickly locate rules that use the Iceberg format.

Unified Catalog¶

Custom Entity Health Calculation with Func: Entities other than the System type can now invoke Func platform functions to calculate health status. A health status filter has also been added to the corresponding entity type list pages.

Infrastructure¶

Database Query Analysis: Added a Query Analysis tab to Databases, providing an aggregated view of normalized queries across instances to help users quickly identify global slow queries.
Network Flow Aggregation Analysis: Added aggregate charts that use time series to display key metrics such as total requests, error rate, average response time, P99 response time, and status code distribution.

Monitoring¶

Data Waiting Window: Monitors can now wait for a specified duration after the scheduled execution time before running, accommodating delayed data reporting scenarios.
Cross-Workspace Alert Policy Search: Backend management now supports searching monitor alert policies across workspaces by member, making notification targets easier to troubleshoot and maintain centrally.

Incident Center¶

Manually Trigger Webhooks: On the incident details page, users can select a configured Webhook channel, add a note, and send incident information to a third-party system. Trigger results are recorded in the incident activity timeline.

Scenarios¶

Dashboard Collaborative Editing Protection: Added concurrent editing notifications and conflict protection to reduce the risk of content being overwritten when multiple users edit a dashboard simultaneously.
City-Level GeoMap: Charts now support city-level GeoMap visualization for more granular geographic data distribution.
Improved Scheduled Report Screenshots: Optimized screenshot generation to reduce blank pages and blank areas.

Synthetic Tests¶

Asynchronous Test Task APIs: Synthetic test task APIs now use asynchronous calls, improving API response efficiency and task execution workflows.

Open API¶

SLS Conversion: OpenAPI now supports SLS data migration scenarios.

Other¶

Expanded Language Support: The platform interface is now available in Japanese and Korean, improving accessibility for users in more regions.

Bug Fixes¶

Fixed incorrect field counts on the Log Export page.
Fixed an issue where data could not be queried after connecting to a Prometheus data source through Func.
Fixed blank screenshots in scheduled report email attachments.
Fixed inaccurate Service counts after filtering Application Service Topology by environment.
Fixed an issue where overseas phone alerts did not work in Managed Edition infrastructure environments.
Fixed an error that occurred when a local user with the Owner role called APIs through Func.
Fixed unexpected query results after navigating from a monitoring event to View All Logs.
Fixed incorrect query results when using BY grouping in log DQL.
Fixed incorrect chart aggregation granularity for a one-day dashboard time range, which caused the number of data points to differ before and after zooming.

August 5, 2026¶

Feature Updates¶

Logs¶

The log export limit has been increased to 1,000,000 entries.
A new export entry has been added to the log context page.
When accessing a non-existent log index, a prompt is now displayed, and the system automatically switches to the first currently loaded index.

Agent Monitoring¶

Page Navigation Optimization: The instance list, Session list, Trace list, and analysis dashboard have been consolidated into a unified entry.
Summary Performance Optimization: Improved the loading performance of the Agent Monitoring summary page.
New OpenCode Agent Type: Agent Monitoring now supports identifying and analyzing OpenCode data.

Incident¶

Support for Assigning Error Tickets by Team: Error tickets can now be assigned directly to teams, not just individuals, facilitating unified triage, assignment, and processing.

Scene Charts¶

The Limit dropdown option for grouped table charts now includes 1000.
Sankey diagrams now support zooming for better browsing of complex network relationships.
DQL view variables now support dynamic log index suffixes, allowing dynamic queries of split log indexes using service name variables, e.g., ${service}_access.

Real User Monitoring (RUM)¶

Agent Auto-integration with RUM: Supports generating integration scripts that allow the Agent to automatically complete the environment configuration for RUM applications, simplifying the integration process.

Plans & Billing¶

New Log Storage Volume Details: The Enterprise Plan billing section now includes "Standard Log Storage Volume Details" and "Long-term Log Storage Volume Details", allowing users to view storage usage per index.

Obsy AI¶

Credit Exhaustion Notification: Supports subscribing to system notifications when Obsy AI credits are exhausted.

Query Workbench¶

The query page has been adjusted to a full-screen popup, and the copy link feature has been removed.
Saved queries now support account-level persistence, with a maximum of 100 saved queries.
New shortcut execution hints: ⌘ + Enter (macOS) and Ctrl + Enter (Windows).

Others¶

Increased Data Forwarding Rule Name Length Limit: The maximum length for data forwarding rule names has been increased from 30 to 64 characters.
New Session Privacy Protection: A "Session Privacy Protection" toggle has been added to Preferences. When enabled, if a user is inactive for more than 30 minutes, the system automatically enters a privacy protection screen.
API Key Rotation Supported: You can now generate a new key value for an existing API Key without needing to delete and recreate it.

Bug Fixes¶

Fixed an issue where scheduled reports were not sent at the specified time, preventing users from receiving them on time.
Fixed an issue with abnormal sorting on the log context page.
Fixed an issue where newly added time series charts in dashboards were not displaying correctly.
Fixed an issue with abnormal data display in dashboard charts.
Fixed an issue where the alert dashboard loaded slowly and displayed inaccurate data.
Fixed an issue where the content of the message field in logs was displayed incompletely or changed abnormally.
Fixed an issue where the sidebar was not displayed correctly when navigating from an iframe-embedded page to log details.
Fixed an issue where cross-space service node jumps in the Service Map were incorrect.
Fixed an issue where related traces could not be displayed correctly in the Incident Center.
Fixed an issue where the RUM Session Replay upload interface in the log module could not correctly parse Multipart forms.
Fixed an issue where key field display items were not automatically refreshed after switching between different indexes.
Fixed an issue where the local disk list displayed abnormally under Chrome version 150.
Fixed an issue where an expression parsing error caused page abnormalities when viewing Profiling data.
Fixed an issue where Profile records failed to open, and JFR file parsing prompted unsupported compression type.

August 1, 2026¶

Billing Adjustments¶

Incident¶

Incident Data Price Reduction: The unit price for error record counting has been adjusted as per the table below. Only the unit price has been adjusted; the billing scope, measurement method, and statistical logic remain unchanged.

Measurement Name: Issue Records
Statistics Scope: Daily count of new associated error data records in the Incident center.
Billing Unit: Per thousand records/day.

Billing Region	Billing Currency	Original Price	New Price
China	CNY	¥1	¥0.14
China	USD	$0.15	$0.02
Overseas	CNY	¥3	¥0.28
Overseas	USD	$0.43	$0.04

July 29, 2026¶

Feature Updates¶

Obsy AI¶

Copilot Query Analysis Upgrade: Optimized query planning and result presentation to improve the accuracy and readability of analysis results, helping users locate issues and understand analysis conclusions faster.
Data Forwarding Query Extension: The OWL tool now supports querying forwarded data, expanding the analyzable data range and providing a more complete data basis for issue diagnosis.
New Copilot Runbook: Supports generating Runbooks based on conversation history or natural language descriptions, enabling Agents to execute repetitive tasks according to preset workflows, improving execution efficiency and result consistency.

Obsy Agent Teams¶

Agent A2A Cross-Space Collaboration: Supports configuring authorization relationships between different Agents, allowing them to call each other within permission boundaries, enabling cross-workspace Agent collaboration.
Runbook Automated Execution: Supports generating Runbooks based on completed tasks or natural language descriptions, enabling Agents to execute repetitive tasks according to preset workflows, improving execution efficiency and result consistency.
Enhanced Long-Task Stability: Optimized context management and interruption recovery mechanisms to reduce information loss and duplicate execution during long tasks, improving the continuity and success rate of complex tasks.
Real-Time Agent Configuration Activation: Configuration changes for Agents, Skills, and MCPs are now synchronized to the runtime environment in real time, reducing wait times and ensuring configuration states match actual capabilities.
Optimized Credit Statistics Query: Improved the query efficiency and stability of credit consumption data, helping users track usage and control resource consumption.

Query Workbench¶

Query Tool Upgrade: The query tool has been fully upgraded to the "Query Workbench", supporting DQL, PromQL manual modes, and multiple query tabs. Users can save frequently used queries, format query statements, copy query links, and export query results to dashboards or CSV files. A new query inspector has been added to view result count, execution time, and status. It also supports generating DQL through natural language descriptions, improving query writing, debugging, and reuse efficiency.

Agent Monitoring¶

New Risk Alert Display: The Session list and call analysis now display risk events and alert levels. Session and Trace overviews also show alert levels. The Trace list now includes the number of risky operations and alert levels, with risk levels marked for dangerous operations, tools, and Skills in Trace details.
New "Display Items" Configuration: Supports customizing the selection of Agent types to display and adjusting the display order of different Agent types on the overview page.
New WorkBuddy Agent Type: Agent Monitoring now supports identifying and analyzing WorkBuddy data.
Optimized the default analysis dashboard template, fixed data display issues, and improved data accuracy.

Incident Center¶

New Time Window Aggregation Mode: Supports aggregating events within a workspace by time window and aggregation keys to generate aggregated incidents, suitable for multi-space event aggregation scenarios.
New Severity Mapping: Supports mapping event severity to incident severity, reusable in aggregation rules and monitor-related incident configurations.
New "Highest Event Severity" Strategy: Aggregation rules can select the highest severity from events already associated with the current incident lifecycle. The default severity order is P0 > P1 > P2 > P3.
Optimized Aggregation Configuration and Filtering: Aggregation keys now support more event fields. The incident list now includes an "Aggregated Incident" identifier and supports filtering by incident type and incident rule.

Synthetic Monitoring¶

New Network Path Test Task: Used to proactively detect target network paths from public or self-hosted test nodes.
- Supports TCP, UDP, and ICMP.
- Supports pre-save testing, scheduled execution, instant testing, and availability judgment.
- Provides a task list, execution records, end-to-end metrics, and per-hop path details.
- Supports viewing partial paths, anonymous nodes, path changes, and standard error messages.

Logs¶

Log Explorer New Pagination Mode: Supports browsing large volumes of log data using pagination, while retaining the original scroll-loading mode, facilitating data browsing and locating.
New Log Index Mapping: Supports matching default log indexes based on fields like app_id, source, service, and their values. When jumping to logs from RUM and other detail pages, the matched index is automatically selected. The log count and log list use the same matching results.
- Supports structured configuration, JSON editing, and format validation.
- Compatible with existing Service matching configurations.
- Automatically falls back to the default index when the matched index is unavailable.
Optimized Log Explorer Interaction: Log indexes can now be switched on the current page or opened in a new tab. "Display Items" and "Quick Filters" will remember the last scroll position. Index search is now triggered by pressing Enter to avoid frequent refreshes during typing.

Unified Catalog¶

The Unified Catalog now supports Entity Status Determination, used to automatically identify the data reporting activity status of hosts and services. In the "Set Judgment Rules" window, you can configure the offline judgment threshold (default 24 hours) and the offline retention period (default 7 days). Manually added entities do not participate in automatic judgment.

Related changes:

Topology View: New online/offline filter, supporting one-click display of only online entities.
Entity List: New "Entity Status" column and status filter.
Entity Details: The top displays a status label. Offline entities can view the offline time, estimated removal time, and support manual activation.

Network Path¶

The Network Path feature now has a dedicated explorer list page, supporting filtering by path status, probe protocol, source, and destination. The list displays key metrics such as path status, end-to-end latency, packet loss rate, and hop count.
The Network Path detail page now includes Historical Probing, displaying probe result distribution over a selected time range using a continuous color band. Clicking a color block allows you to review the path topology and metrics for that time point.

Real User Monitoring (RUM)¶

Enhanced Session Replay Error Messages: Provides clearer status descriptions and handling suggestions for scenarios with no replay data or playback failures.
- Distinguishes between states such as data being generated, incomplete replay data, no replay data, status query anomaly, and replay file fetch anomaly.
- Provides clear reminders for sessions with delayed reporting exceeding 48 hours.
- When an error occurs, the black player is no longer retained; instead, an error page with handling suggestions is displayed.

Data Forwarding¶

Data Forwarding now supports Apache Doris Archive Type: "New Rule" in Data Forwarding now includes Apache Doris as a target archive type, supporting direct forwarding of logs, metrics, and other data to a Doris data warehouse for real-time analysis, further meeting user needs for integrating observability data with mainstream real-time analytical databases.
Optimized Archive File Naming Convention: Nanosecond timestamps are appended to the end of file names, ensuring that multiple files generated within the same millisecond are still in monotonically increasing order, improving compatibility with continuous import (Routine Load) scenarios for Apache Doris and similar data warehouses, reducing the risk of data missing due to file name sorting.

Knowledge Base¶

"Notes" has been renamed to "Knowledge Base". "Tags" have been renamed to "Knowledge Positioning", and "Business Tags" have been renamed to "Business Positioning". Related guidance texts have been updated accordingly.
New Runbook type, supporting automatic generation and invocation of Runbook execution plan documents by Agents.

Management¶

The role management list now includes a "Data Access Rules" column, showing the number of rules associated with each role. Hovering over the count allows you to open a drawer on the current page to directly search, select, and manage associated rules. You can also navigate to the data access management page with the current role filter pre-applied.

Pipeline¶

The maximum character count for Pipeline script content has been increased from 10,240 to 20,480. Existing scripts are not affected, and backward compatibility is maintained.

Bug Fixes¶

Fixed an issue where querying metrics in dashboards prompted invalid operation: "-", while the same metric could be queried normally in Metrics Analysis.
Fixed an issue where functions could not connect to the workspace, causing external data source queries to fail.
Fixed an issue where the jump button for associated Resources was not displayed on the Application Performance detail page.
Fixed an issue where users with an existing billing center account could not create new workspaces.
Fixed an issue where abnormal metrics could not be viewed after enabling host health checks.
Fixed an issue where viewing host data still prompted "no permission" after cross-space authorization.
Fixed an issue where the parsing page was empty after uploading a SourceMap, prompting "No available packages found".
Fixed an issue where newly created test nodes in the backend management did not take effect.
Fixed an issue where the incident aggregation function could not be filtered by tag fields.
Fixed an issue where clicking context on an external index caused a log query error.
Fixed an issue where scheduled report queries did not automatically retry after being throttled, causing some chart generations to fail.
Fixed an issue where the left menu still displayed after hiding it via iframe-embedded pages.
Fixed an issue where trace detail data was not fully displayed after switching Span on the Application Performance detail page.

July 22, 2026¶

Feature Updates¶

Obsy Agent Teams¶

WeCom Message Channel: Added WeCom as a message channel, allowing users to connect Agents in existing WeCom groups for direct alert analysis and incident collaboration, reducing cross-tool switching costs.
My Tasks Management Optimization: Completed tasks will now be automatically archived, and unnecessary tasks can be deleted, reducing historical clutter for a clearer and more focused task list.
Supports Containerized Deployment: Agents now support containerized deployment, with installation instructions and help documentation provided for easy deployment across different infrastructure environments.

Application Performance Monitoring (APM)¶

The call connections in the service upstream/downstream topology now display key call metrics, allowing direct viewing of call volume, error count, and response time between services, helping to quickly identify abnormal calls.

Synthetic Monitoring¶

Optimized error messages for online testing of intranet test tasks. When an intranet address does not support online testing but can run normally after saving the configuration, the page will clearly explain the situation to avoid misinterpreting online test failure as test configuration unavailability.

Logs¶

Optimized log data search: Adjusted related regular expression search capabilities to inverted index queries, improving query performance.
Optimized the display of inverted index-related fields, adjusting the message field to extention_data.

Agent Monitoring¶

Optimized the Agent Monitoring installation process: The installer can now automatically detect the Agent type, simplifying the integration configuration.
Fully upgraded the analysis dashboard, supporting more detailed analysis of Agent calls.

Unified Catalog¶

The "Events" tab on the Entity Details page now includes a Timeline View, displayed alongside the list view. This view overlays change event markers on a status color band, helping users intuitively judge whether changes occurred before and after an incident, shortening the troubleshooting cycle. It is applicable to entity types such as systems, services, hosts, databases, cloud resources, and Kubernetes resources.
Related change: External event monitors now support Marking as Change Events. When enabled, events reported by the monitor are automatically treated as change events and displayed in the timeline of the corresponding entity. This does not affect existing filtering and statistics logic in the Event Center.
The Topology View now includes a "Display Entity Type" filter, supporting quick filtering by entity type (system, service, host, database, cloud resource, etc.). This filter has a union relationship with the left-side filter, allowing users to focus on the asset types of interest.

Infrastructure¶

Database > Top 5 queries now support exporting to dashboards for viewing, as well as exporting to CSV files or copying to the clipboard.

Gateway Integration¶

The OpenTelemetry integration center now supports an Agent auto-integration method. In addition to the original manual environment variable configuration, it now supports generating integration scripts and authorizing the Agent to automatically complete OpenTelemetry configuration in the current environment, simplifying the integration process and lowering the configuration barrier.

Dashboards¶

The dashboard data list now supports switching log clustering analysis results, helping to quickly view and analyze log patterns.
Data precision for units now supports up to 4 decimal places, meeting more granular data viewing needs.

SLO¶

SLOs now support display based on grouping dimension tags even when no events have occurred, facilitating continuous viewing of service target status by business dimension.

Members & Roles¶

Completed the granular permission list for custom roles, aligning permission configuration with actual capabilities.
Optimized the management logic for members, roles, teams, and invitations within a workspace, improving the member collaboration and permission management experience.

Others¶

Optimized default field types: Except for system built-in fields, all other fields are now classified as "Business Fields" by default.
Metrics for both Deployment Plan and Automata plans now support long-term storage, meeting historical metric retention and long-term analysis needs.

Bug Fixes¶

Fixed an issue where the page jittered when switching between data filtering pages.
Fixed an issue where sample debugging information could not be obtained on the metric details page.
Fixed an issue where clicking to edit a view variable resulted in an empty page path, preventing navigation.
Fixed an issue where a network error popup appeared on the Application Performance Monitoring page.
Fixed an issue where exporting application services to a CSV file failed in some scenarios.
Fixed an issue where the Agent Monitoring trend chart displayed abnormally.
Fixed an issue where the application list data was displayed as empty, but data actually existed.
Fixed an issue where query results deviated in some scenarios during multi-index queries.
Fixed an issue where importing a monitoring template in the Management module did not generate the corresponding template.
Fixed an issue where trace data query results between spaces within the same organization were inconsistent with expectations.
Fixed an issue where the data storage cycle display was abnormal under license restrictions.
Fixed an issue where some fields reported errors when creating log indexes in bulk via OpenAPI.
Fixed an issue where the Session search condition was abnormal in user behavior data queries.
Fixed an issue where the substr function calculation result in log view mode was inconsistent with the DQL query result.
Fixed an issue where dashboard configuration reported errors in some scenarios.

July 15, 2026¶

Feature Updates¶

Obsy AI¶

AI Chart Editing: Obsy Copilot now supports adding, editing, and deleting charts within dashboards. Users can select a chart or group and use AI to make adjustments, transforming chart configuration from "manual configuration" to "conversation-driven", significantly improving dashboard editing efficiency.
AI Dashboard Generation Experience Upgrade: Supports displaying view variables during the draft stage, and variables can be updated through conversation, allowing users to adjust analysis parameters during generation and obtain dashboards that better match business scenarios.
AI Feature Entry Expansion: Multiple independent AI feature entries have been added throughout the workspace interface, making it easy for users to quickly invoke AI capabilities in different analysis scenarios like Explorers and dashboards.

Obsy Agent Teams¶

Task Integration Capability: Supports integrating in-space events and external messages into the Agent workflow, allowing Agents to automatically perform task analysis, processing, and result distribution, helping teams convert scattered business events into sustainable, executable intelligent workflows.
- Workspace Integration: Supports selecting an Agent in alert policies. When an event occurs, the Agent can automatically analyze the event and send the analysis results to external message channels like Lark and DingTalk, helping teams complete anomaly diagnosis and collaborative response faster.
- Custom Task Integration: Provides an externally callable Webhook address, supporting user-defined task requirements and execution targets, and sending execution results to external message channels like Lark and DingTalk, making it easy to incorporate tasks from business systems into the Agent's automated processing pipeline.
Channel Integration Experience Upgrade: Adjusted the Channel integration page and the "Create Channel" process, and added configuration documentation entries for channels like DingTalk and Lark, helping users configure and verify message channels more clearly.
My Tasks Experience Upgrade: Completed tasks are automatically archived, reducing list clutter and allowing users to focus on pending items in the Agent.
Enhanced Scheduled Task Capability: Supports more flexible period settings, allowing users to configure automated tasks according to different business rhythms.

Incident Center¶

Added an aggregation rule feature, supporting the merging of alert events from multiple monitors into a single aggregated incident using custom rules, reducing multiple independent incidents and duplicate notifications caused by the same business failure.
[Incident Center > Configuration Management] now includes an aggregation rule configuration entry, supporting the definition of event filtering conditions, aggregation keys, severity calculation strategies, and content display strategies.
Aggregation rules support configuring independent Webhook push addresses, independent of the on-call notification system. When the key status of an aggregated incident changes, the system pushes the full data to the external system.
Aggregated incidents now have two independent fields: system status and process status. The system status is automatically maintained by the platform; the process status supports manual updates or writes via OpenAPI from external systems. The detail page also includes a Webhook push status display and a manual re-push entry.

Infrastructure¶

The Network Connection detail page now includes a Network Path tab, supporting viewing the per-hop network path from client to server, along with per-hop latency and packet loss, helping to quickly locate where network latency or packet loss occurs.

Synthetic Monitoring Support¶

Supports Instant Testing: Supports initiating a one-time instant test for existing test tasks without modifying the task's original enabled or disabled state. Instant test results are written to the Synthetic Monitoring Explorer and can be filtered by trigger method, making it easy to compare availability and response time before and after business releases or changes.
Optimized Test Logic: Optimized the state synchronization logic of test tasks across different nodes and regions. After a task is modified, disabled, or deleted, the system automatically completes the stop state for invalidated regions, reducing task execution deviations caused by residual historical states.

Agent Monitoring¶

Added support for Claude Code, Qoder, and AgentScope types.
Added the obs-agent-connector installer, simplifying the Agent integration process.
Obsy Agent Teams input and output Token consumption is now displayed as credit consumption.

Security Monitoring¶

Added the ABA (Agent Behavior Analysis) monitoring type, specifically designed for AI Agent behavior risk monitoring.
The usage of the ABA type is consistent with existing SIEM and CSPM types. The official rule library includes corresponding rule templates.

Event Center¶

Optimized the display of the Change Event detail page: Long text in "Change Content" is automatically truncated to 256 characters to avoid information overload. The diff comparison is changed from a plain text long string to a structured display, making the changes before and after clearer.

Others¶

Metrics, Logs, and Tracing now include a DataWay data collection guide page, supporting the integration of OpenTelemetry data via the public DataWay data gateway, helping users complete data integration configuration faster.
Added a prompt when exporting the RUM Error Explorer: The exported file will be regenerated based on the current query conditions, so the result may differ from what is currently displayed on the page.
Log export now supports TXT plain text format.
The Service List has been migrated from [Application Performance Monitoring] to [Unified Catalog > Services], and the original entry has been closed.

Deployment Plan Updates¶

Backend management now supports built-in role management capabilities, allowing the creation of role templates and specifying the scope of workspaces where they take effect. The system automatically synchronizes roles to the corresponding workspaces, achieving "define once, take effect everywhere".

Bug Fixes¶

Fixed a spelling error in the URL text on the create workspace page.
Fixed an issue where searching in the Event Center on the US1 site reported missing search index.
Fixed an issue with abnormal display of Agent Monitoring data generated by Obsy Agent Teams.
Fixed an issue where the execution time showed inconsistency between the Infrastructure > Database overview page and the SQL detail page.
Fixed an issue where the credential configuration panel disappeared when selecting AWS S3 as the target for log backup.
Fixed an issue where a monitor did not query some host metrics, triggering a data gap alert.
Fixed an issue where the substr starting position in SLS mode was inconsistent with DQL.
Fixed an issue where selecting the Huawei Cloud RocketMQ detection library in the monitor template library also selected the RocketMQ detection library.
Fixed an issue where "Export All" failed for monitors on the AP1 site.
Fixed an issue where the list view's mute functionality modified DQL and swallowed non-applicable results, causing inconsistent results for the same query in different views.
Fixed an issue where the DQL functions rate_over_count and count produced inconsistent query results under different aggregation windows.
Fixed issues related to Kodo service error logs.
Fixed an issue where clicking on an error in Profiling showed a code stack but could not view details.
Fixed an issue where querying all workspace data at once exceeded the 200-space limit, causing an error.
Fixed an issue where a monitor query returned incorrect results, counting streaming aggregation anomalies and triggering false alerts.
Fixed an issue where the default PV query statement in dashboards did not display correct results in Deployment Plan version 1.130.225.
Fixed an issue where content display in object classification failed.
Fixed an issue where the blacklist could not correctly identify Long Int type conditions, causing filtering to be ineffective.
Fixed an issue where after configuring data access permissions in a workspace, relevant users could not use the Browser filter to filter data on the RUM interface.

July 8, 2026¶

Feature Updates¶

Obsy AI¶

AI Credit Details & Consumption Analysis: Supports viewing the details of Obsy AI credit changes. Using trend charts and heatmaps, it breaks down consumption sources like Agents and Obsy Copilot, helping teams understand AI credit usage and cost changes more intuitively.

Obsy Agent Teams¶

Agent Role Control Optimization: Added a default space role for Agents, used as the default permission configuration for Agents within the space. This helps administrators control Agent permission boundaries more reasonably, reducing the risk of overly broad permissions or misconfiguration.
Agent Intelligence Level Selection: Supports selecting intelligence levels such as Auto, Fast, Standard, and Pro in My Tasks and Automated Tasks, allowing users to flexibly balance response speed, intelligence level, and credit consumption based on task complexity.
Skill Marketplace: Provides an official skill marketplace entry, making it easier for users to discover, install, and manage official skills, reducing the cost of extending and maintaining Agent capabilities, and enhancing Agent usability across different task scenarios.

Agent Monitoring¶

Agent instance names now support Chinese characters, facilitating naming and management by business, team, or scenario.
The plugin update command has been adjusted to an independent entry, no longer placed within the create instance flow, making the creation process simpler.
The Trace detail page call analysis now includes statistics for Model call count, Tool call count, and Skill call count, making it easier to understand the call composition of this Trace.

Unified Catalog¶

Built-in view detail pages now support a time widget, allowing data to be viewed by time range.
The Topology View now supports entity management, allowing modification or deletion of entities.

Incident Center¶

The Incident detail page now includes an "Incident Metrics" tab, supporting viewing of current incident cycle metrics and incident statistics metrics.
The current incident cycle metrics display the trigger, acknowledgment, recovery, and closure timeline for the current cycle, along with three core metrics: Time to Acknowledge (MTTA), Time to Resolve (MTTR), and Current Duration.
Incident statistics metrics are calculated cumulatively from the first trigger, displaying the first trigger time, last trigger time, total reopen count, average MTTA, and average MTTR.
When an incident reopens, the current cycle metrics only display the latest cycle data, while the statistics metrics continue to accumulate.

SLO¶

SLOs now support a grouping dimension feature, allowing tracking of compliance rates for different project, service, and env values within the same SLO, eliminating the need to create a separate SLO for each business line.
When creating or editing an SLO, a "Grouping Dimension" configuration area is now available below the SLI section. After selecting an SLI monitor, the system automatically reads the dimension values actually present in the monitor's data over the last 7 days.
Below the title on the SLO detail side panel, a dimension filter row has been added, supporting switching between different dimension value combinations to view their compliance rates.
When configuring an SLO chart in a scene dashboard, a "Filter Dimension" row has been added to the query area. After selecting specific dimension values, the chart displays the corresponding compliance rate in real time.

Note

Each dimension combination is an independent calculation line and does not support cross-dimension aggregation. To view the compliance rate at the project level, you need to create a separate SLO grouped by project. The behavior of old SLOs without configured grouping dimensions remains unchanged, ensuring full backward compatibility.

Notification Targets¶

The Webhook notification target now includes a Custom Body Template configuration option. When enabled, users can freely write the request body content and reference alert context variables, meeting the custom format requirements for pushing alert data to downstream systems like Kafka. When not configured, the Webhook maintains the original fixed format for backward compatibility.

Dashboards¶

The simple query mode now supports multi-select for the same aggregation metric, allowing multiple metrics to be selected simultaneously in the same query line, aligning with the multi-metric query capability of a single statement in DQL mode.

Management¶

The Role IP Allowlist rule lines for Login IP Allowlist and OpenAPI Access IP Allowlist now include a remark input box, supporting custom descriptions for each role rule.
OpenAPI now includes a member personal API Key permission switch interface: POST /api/v1/workspace/member/personal_api_key/enable/modify, supporting batch enabling or disabling of member's personal API Key usage permissions via API Key.
The caller must have the workspace.personalAPIKeyManage permission. Only API Keys with the Owner or Workspace Administrator identity can call this interface.

Data Forwarding¶

The forwarding rule list page now includes quick filter options, supporting quick filtering and locating rules based on data type, storage format, and archive type.
The object storage rule configuration page now includes a "Data Wait Time" configuration. Data is first written to a local temporary file. When the file size reaches 256 MB or the wait time reaches the configured value, the system automatically compresses and uploads.

Generate Metrics¶

The frequency configuration for generating metrics now includes preset options like 30 minutes, 1 hour, 6 hours, 12 hours, and 24 hours.
Supports custom input of generation frequency within 24 hours, meeting more flexible metric calculation cycle configuration needs.

July 1, 2026¶

Feature Updates¶

Notes¶

The Notes feature has been upgraded to an independent first-level module, with the entry moved to the top navigation bar.
The Notes page has been redesigned with a three-column layout, supporting tag/business tag filtering, quick filtering, multiple tabs, and state retention.
Notes support visibility, global tags, business tags, and other attribute management. The body uses Markdown storage, and a new Chart Block syntax has been added for inserting renderable charts into Markdown notes, supporting 6 chart types: time series, summary, table, bar, histogram, and pie.
The Unified Catalog entity detail page now includes a "Notes" association module, which can automatically aggregate and display related notes matching business tags, and synchronize the favorite status.

Obsy AI¶

Optimized Note Generation: Obsy AI-generated notes now adapt to the new note format, supporting direct editing after generation.
Obsy AI Stability Optimization: Optimized the task execution and result generation chain for Obsy AI, improving feature stability and overall usability.
AI Automated Dashboard Creation Optimization: Supports directly generating dashboards with view variables, improving dashboard configuration efficiency and subsequent reuse experience.

Obsy Agent Teams¶

Agent Configuration Experience Optimization: Unified the global configuration and personal configuration entries, and added installation and usage instructions for CLI, Obsy MCP services.
Agent Information Display Optimization: Optimized the Agent creation guidance and information display experience, improving Agent configuration and viewing efficiency.
Create Task Experience Optimization: Optimized the task creation process to help users start Agent tasks more efficiently.
Enhanced Service Security: Improved the security of Agent service communication and deployment, supporting safe use in more complex and strict deployment environments.

Plans & Billing¶

Obsy AI Credit Pre-purchase: Obsy AI credits support pre-purchase top-up. When the plan credits are exhausted, pre-purchased credits will be used. Pre-purchased credits are valid for a long period.
Pro / Team Plan Credit Adjustment: Pro and Team plans now include more Obsy AI credits.
Obsy AI Credit Notifications: Added notifications for insufficient credit balance and credit exhaustion to prevent service interruption.
Log long-term storage billing is calculated based on the raw storage volume of log data within the current total long-term storage period, minus the raw storage volume within the standard storage period (per GB/day). That is, only the log data exceeding the standard storage duration is billed.

Application Performance Monitoring (APM)¶

The service list and service map now prioritize displaying service aliases. Users can set service aliases in [Unified Catalog] to distinguish different business services with more recognizable names. If no alias is set, the original service name is still displayed.

Logs¶

Log indexes now support cross-page sorting adjustments, allowing one or more indexes to be moved to a specified position, suitable for scenarios with many indexes requiring fine-grained matching order adjustments.
Upgraded the log index matching condition configuration, supporting clearer combination of multiple matching rules and a more intuitive display of the final matching logic. Added Advanced Matching billing logic, while retaining the free standard matching.

Monitoring¶

Notification targets now include 4 new types: Voice Call (China), Voice Call (International), SMS (China), SMS (International).
The guance.com overseas site now includes an international zone partition, supporting voice calls and SMS notifications with international area codes (+1, +44, +81, etc.).
The guance.one all sites now include a China zone partition, supporting voice calls and SMS notifications for +86 China mainland numbers.
Existing notification target types remain compatible and do not require reconfiguration.

Note

The China zone only supports +86 China mainland numbers; international area codes cannot receive notifications. The international zone only supports international area codes; +86 numbers cannot receive notifications.

Agent Monitoring¶

Added support for the Codex type, optimized the type display order and field specifications, and added time widget refresh capabilities for easier viewing of the latest monitoring data.

Generate Metrics¶

Added a "Data Wait Window", supporting configuration of delay durations like 0 seconds, 1 minute, 5 minutes, etc., to specify the wait time for the computation task after the task window ends. Adding a data wait time can accommodate client data reporting delays, improving the accuracy of generated metrics.

Open API¶

Supports adding, deleting, and querying custom templates via Open API.

Others¶

Optimized the dashboard view variable hiding feature. In the dashboard configuration > View Variable list, you can toggle the visibility of each variable using the eye icon on the right. You can also set it via the "Hide" switch in the variable editing page. The two states are synchronized bidirectionally in real time.
Infrastructure > Pods > Hexagon chart now includes "CPU Usage (Standardized)" as a fill field option.
Unified Catalog > Configuration Health Window now includes a close button, supporting one-click disabling of health calculation.
Integration > Extensions now includes Func application status display, supporting manual deletion of offline applications.

Integrations¶

Updated Flink dashboards and monitors.
Added Huawei Cloud GminiDB Influx integration.
Added Huawei Cloud GminiDB Redis integration.
Added Volcengine Cloud Billing integration.

Enterprise Plan Updates¶

Supports log long-term storage:

Billing logic: Log long-term storage billing is calculated based on the raw storage volume of log data within the current total long-term storage period, minus the raw storage volume within the standard storage period (per GB/day). That is, only the log data exceeding the standard storage duration is billed ("Total Storage - Standard Storage").

Bug Fixes¶

Fixed an issue where some upstream call relationships were not correctly displayed when loading the APM service map for the first time.
Fixed an issue where data retrieval failed due to query interface anomalies in the User Analysis module.
Fixed an issue with abnormal filter query behavior in the dashboard panel.
Fixed an issue where OpenSearch did not automatically create a new index after a single index reached the document limit.
Fixed an issue where clicking to view trace analysis across workspaces incorrectly redirected to the SSO login page.
Fixed an issue where modifying the name of a Synthetic Test self-hosted node caused an abnormal node_name return and a frontend error.
Fixed an issue where the numerical precision in dashboards was not correctly rounded, causing display deviations.
Fixed an issue where monitoring charts for CPU, disk, etc., could not be displayed after integrating AWS MySQL databases.
Fixed an issue where the "View Related Logs" link in alert notifications could not navigate to the Log Explorer.
Fixed an issue where some scheduled report tasks were not generated as expected.
Fixed an issue where the Side Panel Index and Banner status did not refresh synchronously after Data Access rule changes.
Fixed an issue with duplicate DataWay displays on the backend management page.
Fixed an issue with abnormal display of the test explorer table header, abnormal field management search, and missing internationalization translations.
Optimized dashboard query performance and page rendering efficiency.
Fixed issues related to kodo-x error logs discovered through self-observability, improving platform stability.

June 24, 2026¶

Feature Updates¶

Logs¶

Optimized the JSON expansion level experience in stacked log mode: Supports choosing not to expand or fully expand in preferences, improving the efficiency of browsing multi-layer structures, locating fields, and troubleshooting.

Application Performance Monitoring (APM)¶

Optimized the service map display: The service map now includes a "Show only services matching filter criteria" toggle. When enabled, only service nodes matching the current filter criteria are displayed, hiding non-matching nodes.
The resource topology now displays the total number of requests. Added a legend to help users understand the request statistics scope.

Synthetic Monitoring¶

Test billing has been separated from log billing, supporting independent statistics and management of test billing.
Test storage duration can now be set independently.
Adjusted the billing rules for self-hosted test nodes: When using a self-hosted test node, the billing count is calculated as 1/10 of a public test node. That is, each browser test execution counts as 1, and each execution of other test types counts as 0.1.

Queries¶

Monitor and dashboard queries now include accelerator field configuration guidance. When the selected field is not configured as an accelerator field, a prompt is displayed below the query line, guiding the user to enable query acceleration in the corresponding module's global configuration page.

Dashboards¶

Optimized the query interaction for "Summary Charts", supporting the addition of multiple queries. By default, only the most recently added query is enabled; other queries are automatically disabled. Users can manually switch and enable a specific query.

Management¶

Workspace Settings > Security: Added a global toggle for "Data Access Scope Restriction". When enabled, members who do not match any data access rules in the current workspace have no permission to view data by default. Only the Owner and Administrator roles can operate this toggle and are not affected by it.
Workspace Settings > Security: In the configuration popup for Login IP Allowlist and OpenAPI Access IP Allowlist, a "Role IP Allowlist" configuration block has been added, supporting the configuration of IP allowlist rules by role dimension. Users who do not match any role rule will fall back to the global IP allowlist judgment.
Data Forwarding: The rule list now includes a "Show only abnormal rules" filter option, supporting one-click viewing of abnormal rules.
Metrics Management: Added a "Source" display column for metrics, making it easier to identify the metric source.
Added audit records for enabling/disabling member personal API Key permissions, supporting tracking of permission change sources and affected members.

Plans & Billing¶

Central Pipeline Data Processing Traffic Price Reduction: The price for central Pipeline data processing traffic has been reduced by approximately 50%. This adjustment only involves the unit price. The billing scope, measurement method, and statistical logic remain unchanged. It is still billed based on the daily raw data volume that hits the central Pipeline.

Site	Settlement Currency	Price Details
China	CNY	¥0.05
China	USD	$0.007
Hong Kong & Global	CNY	¥0.084
Hong Kong & Global	USD	$0.012

Open API¶

Supports modifying the storage policy of a single Measurement via Open API.

OWL¶

Added same-org cross-workspace query capability, supporting the same_org_workspace_uuids parameter and owl workspace same-org list command.
Added mdsearch document retrieval tool, supporting searching, reading documents, and browsing directories.
Added data and analysis-related tools: owl.slo.list, owl.data.simple_query / owl.data.simple_query_file, owl.logging.cluster_task.create / .get, owl.billing.list.
Added OWL CLI workspace and configuration capability documentation, including owl workspace related commands, OWL_API_KEY, and OWL_DIR.

June 17, 2026¶

Feature Updates¶

API Key Security System Comprehensive Upgrade¶

To address the increasingly complex data security challenges of the AI era, Guance has undergone a comprehensive security upgrade for its API Key system.

Core improvements:

Self-Destruct Mechanism: Workspace-level and personal-level API Keys now support a self-destruct mechanism. The key is only fully displayed at creation time. After the page is closed, the plaintext cannot be viewed again, reducing the risk of key leakage.
Enhanced Signature Authentication: OpenAPI and WebSocket interfaces now fully support AK/SK signature authentication, replacing traditional plaintext transmission to improve the security of API calls.
API Key ID Removal: The API Key ID is no longer displayed on the frontend and is not generated during creation, simplifying user cognition while reducing the exposure of sensitive information.
Func Compatibility Guarantee: The new Func version is compatible with the old Guance, and the old Func version is compatible with the new Guance, ensuring a smooth transition.

Obsy Agent Teams¶

Obsy Agent Teams now supports enabling self-observability, allowing viewing of call trend details in Agent Monitoring.

Plans & Billing¶

The DataKit self-monitoring metric dk is no longer included in metric billing. It is enabled by default upon installation, allowing users to perform troubleshooting and observational analysis at any time.

Logs¶

Added a "Full-Text Index Fields" configuration in the advanced options of log indexes, supporting two modes:
- Only message index field (default): Full-text search only matches in the original message field. The system retains the original log content.
- Full-line index: Full-text search matches across all business fields. The system no longer saves the original message field. Extracted business fields are written to variant for retrieval.
Log Explorer:
- Supports @field automatic mapping: When using @key:value queries, the system first tries to match top-level fields. If not found, it automatically falls back to searching within the message JSON field. When using the complete message@key:value syntax, the original logic of directly querying the message JSON is maintained.
- Optimized query acceleration configuration interaction: The query acceleration configuration has removed the acceleration status toggle, replaced by a delete operation. A new entry for adding accelerator fields has been added, supporting dropdown selection or manual input. The maximum number of accelerator fields per index is 200.

Unified Catalog¶

The Topology View now supports theme color adaptation.
On the Entity Details page, the entity list supports clicking to open a side panel to view associated entity details, and supports returning to the previous level.
Optimized the service entity synchronization logic: tracing_service is no longer automatically written to the Unified Catalog.

Incident Center¶

On-Call Management Interaction Upgrade:

The on-call home page now supports a Card / Calendar dual-view toggle. The card view supports filtering by All / Mine and search, with the current on-call person automatically pinned to the top. The calendar view supports viewing schedules by month, with expansion for days with more than 4 entries.
On-call configuration optimization: The original "Time" has been renamed to "Working Time", and the original "Effective Time" has been renamed to "Effective Scope". The rotation cycle now includes a custom option, supporting settings by 1–31 days or 1–4 weeks.
Supports creating Temporary On-Call Assignments. After specifying the substitute and time, the substitute will receive incident alerts during the substitution period.

Agent Monitoring¶

The overview page has been restructured into a dual-view for Agent / LLM, supporting viewing integration and running status by type. Added KPI overviews such as request count, Tokens, and latency, along with the most recently active instances.
The Explorer and Analysis Dashboard now support switching between Agent / LLM states. In Agent state, filtering is supported by Agent type and instance level. In LLM state, filtering by application is supported.
Updated the Agent Integration Configuration guide, supporting selecting the integration type, entering the instance name, and generating the integration code. Provides connection verification and plugin update guidance.

Scenes¶

View variables now include an "External Function" query method, supporting the invocation of functions registered in the workspace to dynamically generate variable candidate values.

Event Center¶

Optimized the query condition for navigation from the event detail page: When clicking "View All Traces" from the event detail page, if the event dimension value is <nil> or empty, the system no longer brings it in as a normal equal condition. Instead, it automatically converts to a NOT EXISTS syntax, avoiding empty query results. The navigation logic for normal dimension values remains unchanged.

Metrics¶

Metrics Analysis: The page now includes a "Copy Link" button, supporting sharing and navigation via link.

Monitoring¶

Mute Rules: Added a "Simultaneously Mute Incidents" configuration option, supporting the suppression of incident notifications while muting events. When configuring mute based on monitoring rules, monitor tags, or custom scope, this option can be enabled. When enabled, incidents will not trigger any notifications during the rule's effective period.
Optimized the display of alert policy notification target configuration.

Synthetic Monitoring¶

API Tests now include an SSL Protocol Type, supporting periodic detection of the TLS/SSL certificate status of target domains. You can configure availability judgment conditions such as certificate remaining days and response time, and support initiating TLS handshake tests from test nodes to verify connectivity.
The task page now includes an Overview View: Supports aggregated display by test type, using a treemap to intuitively show the scale and health status of each group of tasks. Clicking down allows viewing the latest test result of a single task and supports one-click navigation to the Explorer for analysis.

APM¶

Trace Query: Added the same_org_workspace_uuids query parameter, supporting cross-workspace querying of trace data within the same organization by trace_id via OpenAPI and AI API interfaces.

Others¶

Now supports Elasticsearch Kerberos authentication, meeting the integration needs of ES clusters using Kerberos security authentication mechanisms.
The monitor configuration page and dashboard chart query page now include a SLS to DQL translation capability for log query lines. When an SLS query statement is entered, the system automatically translates it to DQL and displays the final execution statement. Preview, save, and alert execution all use the translated DQL. The translation result supports switching to DQL edit mode for further modification, helping users migrating from Alibaba Cloud SLS reduce query statement rewriting costs.
Data Forwarding:
- The list now includes a "Display Columns" configuration capability, supporting controlling the display state of list fields on demand. "Forwarding Rule" is a fixed display column, while other fields can be configured to show or hide via a toggle.
- Added a "Storage Path" column.

June 10, 2026¶

Feature Updates¶

Plans & Billing¶

Obsy AI Plans & Billing Model Upgrade: Obsy AI now offers four plan subscriptions: Basic, Pro, Team, and Enterprise. Supports monthly subscription with auto-renewal. Each plan includes corresponding Credits. The Basic plan resets monthly. Paid plans allocate Credits periodically, valid only within the current subscription cycle (unused Credits expire at the end of the cycle and do not accumulate). Pro and Team plan allowances have been increased. Added plan upgrade (immediate effect, with price difference and Credits adjustment) and downgrade (effective next subscription cycle) capabilities. Credits consumption first deducts plan-included Credits. After plan Credits are exhausted or expired, pre-purchased Credits are used.
Security Check Rules & Programmable Monitor Billing Adjustment: The billing method for Security Check rules (SIEM/CSPM) and Programmable Monitors has been changed from per compressed scan traffic to per detection count. Each detection counts as 100 Triggers. The original scan data volume is no longer used as the billing basis.

External API¶

ExternalAPI now supports read/write account separation: Added a read-only ExternalAPI configuration to achieve read/write permission isolation.

OWL¶

OWL CLI Multi-Workspace Management: The OWL CLI now includes the workspace command for managing multiple workspaces.
- owl workspace list retrieves the list of available workspaces under the current account.
- owl workspace use <workspace_uuid>: Switches to a specified workspace. It automatically fetches the access key from the cloud and writes it to the local configuration. Subsequent commands like owl sync and owl exec will use this workspace identity. The auto-installation process is also compatible with multi-workspace configurations.
OWL Data Query Sampling Information: The owl.data.query result now includes the sampling status and sampling rate information, helping to determine whether the query result is complete.

Management¶

Added "Workspace IP Allowlist": Supports configuring an IP allowlist at the workspace level to restrict source IPs for console access and OpenAPI / API Key calls. Supports CIDR format. Requests from non-allowlisted sources will be rejected, and audit logs will be recorded.
Personal API Key Permission Control: The Member Management list now includes a "Personal API Key" display item, supporting workspace administrators to batch control whether members have the permission to generate personal API Keys.
Log visibility delay now supports index-level configuration: Different delay times can be set for different indexes. When querying multiple indexes, the strictest delay time applies automatically. Indexes without a configured delay are unrestricted.

APM¶

Service > Performance Metrics: Added a total request count column, showing the total number of requests and a trend mini-chart for each service within the selected query time range. Also supports customizing whether to display this in the display items.

Monitoring¶

Alert Notification Linked with AI Analysis: Optimized the AI analysis entry in alert notifications. Clicking to navigate to the event detail page now automatically expands the Obsy AI panel and triggers analysis of the current alert event. Supports automatically restoring the analysis state after login if not logged in.

Synthetic Monitoring¶

Added the ability to drag column headers in the task list.
Added a custom display item button in the task list, supporting custom display behavior for tags.
Quick Filter > Tags: Added a "High-Frequency Tags" module, displaying the top 10 most used tags.

June 3, 2026¶

Feature Updates¶

Obsy AI¶

Added support for generating notes: Supports organizing the current conversation content into a platform note with one click. After completing tasks like troubleshooting or log analysis, the AI can actively suggest or generate a structured note based on user instructions. After successful creation, the user can directly navigate to view it.

Unified Catalog¶

Entity Type Management Upgrade:
- The list has been changed from a popup to an independent page, supporting search by name and display name. When creating/editing, you can customize the Data Schema, display columns, and associated views via YAML in "Advanced Configuration".
- Enhanced display column configuration: Supports setting fixed columns and default hidden columns. User personal column preferences take precedence over type default configuration.
- Log association views now support default indexes: The index field can be configured via YAML, and the entity detail page will automatically match the corresponding log data.
- Added K8s Service and Deployment auto-write entity types.
- Data Schema extension: Official types support appending custom attributes. Custom types support custom fields, source, required, enumeration, and validation rules.
Optimized the Topology View Token expiration prompt: When the Token expires due to prolonged inactivity in the console, a one-click refresh for recovery is supported.

Scenes¶

View variable optimization: Only variables actually referenced in the current chart query statement are displayed. Unreferenced variables are automatically hidden, reducing irrelevant variable interference and improving the data analysis experience.
Stacked charts now support displaying the sum: Line charts and bar charts in stacked mode now include a "Show Sum" toggle (default off). When enabled, the tooltip displays the total value of all visible queries at the current time point at the top.

Incident Center¶

Supports cross-space viewing of incidents. The incident list now includes a space range selector.
- Related change: The data scope for Cross-Space Authorization now includes "Incidents".
On-call configuration now includes an "Auto-Acknowledge" toggle. When enabled, if an incident matches only one on-call person, that member automatically becomes the assignee, and the incident status is updated to "Processing".
Added a "Current On-Call Person" option in Level 0 rules, referring to the on-call person currently effective in real time after the on-call strategy rotation.

Agent Monitoring¶

Added support for creating Agent Monitoring applications via Hermes integration.

APM¶

The service map now supports space scope switching: By default, only service nodes and call relationships within the current workspace are displayed. When cross-space authorization exists, spaces under the same site can be multi-selected. Cross-site selection is limited to single selection. Search, filtering, and metric statistics are calculated based on the selected scope in real time.
- Related change: SaaS version supports administrators configuring the "authorized workspace → authorized workspace" ServiceMap data viewing permission, supporting cancellation of authorization at any time with immediate effect. The Deployment Plan version supports cross-space authorization.

Synthetic Monitoring¶

Added support for Browser Tests, which can record page operations via a browser plugin or upload YAML scripts to simulate real browser access paths, monitoring Web application availability.
- Related change: The test execution detail page displays key performance metrics such as TTFP, LCP, CLS, step-by-step execution results, and retry records.

Infrastructure¶

Database:
1. Added AI optimization suggestion entry on the Query detail page: Clicking the button in the upper right corner of the SQL statement area will invoke the Obsy AI side panel and automatically execute Query optimization analysis.
2. The load trend charts from the original query and session modules have been consolidated into a single chart, supporting grouping by dimensions such as wait type, database, user, client, and application. The default display shows load distribution by wait type.

Explorers¶

Quick filters now support coexistence of global configuration and personal customization. When an administrator updates global filter options, all users can see the latest configuration after refreshing the page. Personal customized filter options are not affected.
Global configuration: Added top field configuration. Supports customizing the fields displayed at the top of the log in stacked mode. You can set a global default configuration or configure dedicated top fields for different indexes. Indexes without a dedicated configuration will automatically use the global default configuration.
Quick filters: Supports null (no data) filtering. An administrator can enable this in the global configuration. When enabled, a "No Data" option is displayed at the bottom of each filter group.

Plans & Billing¶

Usage Analysis: Data ingestion traffic now supports splitting by type, providing two views: categorized and aggregated. The view supports switching between traffic/count units and line/stacked chart display. The time range can be freely zoomed. In the aggregated view, clicking on one or more metrics allows for aggregated analysis, or you can view a single metric separately.

Management¶

Data Authorization:
- Supports cross-site workspace data access authorization via authorization packages, including SaaS Commercial > SaaS Commercial, SaaS Commercial > Deployment Plan, Deployment Plan > Deployment Plan.
- The authorization list now includes status indicators for "Pending Import", "Active", and "Expired", making it easier to track the authorization progress.
The personal account page now includes a quick entry for language switching.

Bug Fixes¶

Fixed an issue with abnormal display of the Topology View in the Unified Catalog.
Fixed an issue where charts in dashboards could not display specific metrics during editing.
Fixed an issue where the large query mode returned no data, but the original query returned data.
Fixed an issue with abnormal display of .NET performance analysis data.
Fixed an issue with excessively long OIDC response times.

May 27, 2026¶

Feature Updates¶

Obsy AI¶

History Dialogues: Added a history dialogue entry, supporting viewing, continuing, renaming, and deleting past conversation records. History dialogues are sorted by update time in descending order. When continuing to ask questions, the original conversation context is retained. History is saved for up to 30 days.
Answer content now supports structured display. Compared to the original plain text Markdown display, the new display format automatically presents different visualization components based on the content type, including summary information, analysis conclusions, list steps, table data, risk warnings, etc., improving the readability and information hierarchy of the results.

Time selector interaction optimization: The time selectors in Data Forwarding > Data Viewing and Log Long-term Query now support cross-day timeline drag adjustment.
In the above scenarios, a query interval preview is now displayed at the bottom of the time selection panel. When dragging the timeline or using the scroll wheel to adjust the start and end times, the currently selected time range and total duration are displayed in real time.

Management¶

Lab: Optimized the SLS mode query interaction. After executing a query, the corresponding DQL statement is displayed below the input box. Directly editing this DQL and executing it will automatically switch to DQL mode. When using the left-side quick filter in SLS mode, a confirmation prompt is added. Clicking "Continue" will synchronize the filter conditions and switch to DQL mode.
Preferences: Added theme color selection, including Default Blue, Ocean Blue, Green, etc.

Scenes¶

Charts > Unit: Added Indonesian Rupiah (Rp) as a currency unit.

APM¶

Trace Detail Page: The Span list and waterfall chart list now support selecting to display the full URL.

Synthetic Monitoring¶

TCP Test: Optimized the default timeout value for each hop in route tracing to 500ms, covering intermediate node latency in more real network scenarios, preventing normal packets from being misjudged as timeouts.

May 20, 2026¶

Feature Updates¶

Obsy AI¶

Added a Obsy AI subscription entry. The workspace Owner can enable the Obsy AI intelligent service here. After activation, advanced capabilities are billed based on actual AI Credits consumption.
Added a Obsy Agent Teams activation entry at the top of the Obsy AI drawer homepage.
Adjusted the Obsy AI billing model to: Basic Q&A Free + Advanced Capability Free Trial Quota + Agent Monthly Subscription Fee + Credits Consumption + Dedicated Voucher Offset + Pay-as-you-go for Excess:
- Obsy AI Intelligent Service:
  - Basic Q&A: Free to use, does not consume the free trial quota or AI Credits.
  - Obsy AI Intelligent Service Not Enabled: The workspace provides 50 free advanced capability trials per month, shared by all members in the workspace. The quota resets each calendar month and does not roll over.
  - Obsy AI Intelligent Service Enabled: Advanced capabilities are billed based on actual AI Credits consumption. The 50 free trials are no longer consumed. Billing first uses the earliest expiring dedicated vouchers to offset Credits consumption. If vouchers are insufficient, the remaining cost is billed at a fixed unit price on a pay-as-you-go basis, settled daily.
  - Service Disabled: Reverts to 50 free advanced capability trials per month. Basic Q&A remains free.
- Obsy Agent Teams: Uses a monthly subscription model.

Related changes:

Incident Center > Configuration Management: Added an Incident Auto-Analysis feature. When enabled, for each new incident generated, the system automatically invokes Obsy AI for analysis and appends the result to the incident collaboration record. This feature is only available for workspaces that have enabled the Obsy AI Intelligent Service. The toggle is grayed out and disabled if the service is not enabled.
The original Issue > Issue Auto-Analysis feature has been closed and is no longer supported.

Plans & Billing¶

Logs: Added a "Long-term Storage" billing item, calculated based on the raw storage volume of data within the long-term storage period (including standard storage period data).
Metrics: After enabling Long-term Storage, the system automatically reduces precision to compress data points and save space, integrated with normal metric billing, without separate billing. Single Measurements support independent storage policy configuration, which will no longer follow the global policy after customization. This feature is only available for Commercial Plan workspaces billed by data points per month. Workspaces billed by time series are not supported yet.

Unified Catalog¶

System Entity List:
1. Added support for List and Card dual-view switching. The list mode supports sorting by health status and expanding to view component entity details. The card mode displays system summaries such as total entity count and last alert time.
2. Added Health display. Health supports two calculation methods: system type-level default configuration and individual system custom configuration.

Scenes¶

Added a new chart type "Dashboard List". As a navigation chart, it is used to display dynamic links pointing to other dashboards within the current dashboard.
Resource Topology Chart: Added support for linked view variables. Clicking a node in the topology chart can add the corresponding resource or service as a view variable filter value with one click, enabling linked analysis with other charts on the dashboard.

Log Explorer¶

Added a "Standard Query" and "Long-term Query" switch entry. Standard Query supports real-time retrieval for any time range. Long-term Query is suitable for historical log review, with a maximum single query time range of 24 hours. When the query range exceeds the standard storage duration, the system will automatically prompt and guide the user to switch the query mode.

Incident Center¶

The list page supports batch acknowledging incidents.

Infrastructure¶

Database List Page: The database list page now includes database type icons, supporting QPS and load trend thumbnail display.

Synthetic Monitoring¶

Advanced Settings > Request Settings: Added HTTP version configuration, supporting selection of HTTP/1.1 (default), HTTP/2, or HTTP/3. When selecting HTTP/3, proxy configuration will not be supported simultaneously.

Management¶

API Keys: Added Security Display Policy. When this policy is enabled, newly created API Keys are only displayed in plaintext once at creation time. After closing the page, they cannot be viewed again.

Others¶

The entire platform's phone number input now supports selecting a country/region code. Scenarios such as registration, login, notification targets, alert configuration, and personal center can all use the area code selector to fill in the phone number of the corresponding country. Existing China mainland phone numbers default to the +86 area code, which does not affect existing users.
For users in the Hong Kong and Global sites, a new phone number entry field has been added. If a phone number is not provided during registration, users need to complete it in Account Settings to avoid affecting the receipt of SMS and IVR alert notifications.

Bug Fixes¶

Fixed an issue where logs associated with traces and infrastructure were not displayed on the first click of a log.
Fixed an issue where the email sending function failed when switching the registration page to Indonesian.
Fixed an issue where the language of the intelligent analysis results was inconsistent with expectations.
Fixed an issue with abnormal log clustering analysis data for OpenSearch versions.
Fixed an issue where variable filtering was not successful in variable data source queries.
Fixed an issue where the corresponding unit of the metric was not displayed when configuring monitoring thresholds.
Fixed an issue where query results were inaccurate when sub-queries carried having filters in eval mode.
Fixed an issue where the Incident interface could not display all content and the page could not scroll down.
Fixed an issue where all components in the Global site dashboard view displayed a count of zero, but the preview in edit mode showed correct data.
Fixed an issue where some items were missing in the billing statistics interface.
Fixed an issue where dashboard queries in scheduled reports failed and were delayed.
Fixed an issue where charts in dashboard scheduled reports displayed query timeouts.
Fixed an issue where dashboard scheduled reports were not sent as expected.
Fixed an issue where the stack trace information in the Error Explorer was inconsistent with the content displayed by the parsing tool.
Fixed an issue where querying logs in ascending order using the Open API did not correctly return the cursor time value.
Fixed an issue where the interface loading prompt display was not aesthetically pleasing.

May 13, 2026¶

Feature Updates¶

Dedicated Compute¶

Added billing item: Dedicated Compute. Supports purchasing dedicated compute resources at the workspace level. After purchase, all query requests for that space will use dedicated compute nodes, no longer sharing query resources with other workspaces, greatly improving the query experience. Available in multiple sizes: Large, XLarge, 4XLarge, 8XLarge. Billed monthly.

Agent Monitoring¶

The original LLM Monitoring has been officially upgraded to Agent Monitoring, with the following updates:
1. Application List: Supports creating Agent / LLM application types, displaying core metrics such as call count, Tokens, and latency. Supports editing and deleting.
2. Explorer: Added Session List and Trace List, supporting drill-down to view waterfall charts, call details, and Model / Skill / Tool call proportions.
3. Analysis Dashboard: Agent applications display four analysis modules: Requests, Models, Skills, and Tools. Supports filtering by Agent application.

Management¶

Added Lab: Used to centrally manage experimental features. All experimental features are disabled by default and can be manually enabled on demand. They only take effect in the current workspace after being enabled.
- Added "SLS Language Parser" experimental feature under the Lab: When enabled, the Log Explorer supports using standard SLS query syntax for data retrieval. The query entry is upgraded to "Advanced Query", allowing flexible switching between DQL mode and SLS mode.
Member Management:
- Member selection optimization: All member selection dropdowns (monitor permissions, mute rules, alert policies, notification targets, scene visibility, Incident/Incident Center, etc.) now support distinguishing members with the same name. When the list contains members with the same name, the email is automatically displayed. If the name and email are both the same, the account type is further displayed. Selected member tags follow the same rule.
- Member management list optimization: The member management table now includes an "Account Type" column. A "Settings" button has been added above the list, supporting custom display columns. The settings are also applied to the member export file.
API Keys Management > List: Added a "Last Call Time" display column, supporting sorting the list by time.
Data Storage Policy: Added support for 60-day and 90-day configurations.

DQL¶

DQL Query: Supports CASE WHEN conditional expressions. Used to select different values based on conditions in a query. Typically used with aggregate functions for conditional counting, conditional summing, or normalizing fields by condition. Basic syntax:

CASE
  WHEN condition THEN value
  [WHEN condition THEN value ...]
  [ELSE default_value]
END

APM¶

Trace Explorer: When clicking to sort by the "Time" or "Duration" field, if the field has a high cardinality, the system now displays a popup prompt (Due to the high cardinality of this field, the query wait time may be long. It is recommended to narrow the query scope before sorting). You can choose to cancel or continue sorting.

Deployment Plan¶

Custom storage days now include 60-day and 90-day storage policies, with the storage upper limit extended to 3600 days.

April 29, 2026¶

Feature Updates¶

Scenes¶

Histogram > Chart Query: Supports PromQL queries.

Management¶

Workspace Settings > Advanced Settings: Added a "Log Visibility Delay" toggle. Supports setting a unified delay visibility window for log queries. When enabled, all users (including workspace owners and administrators) can only query log data before the delay window. Log data newly written within the window period is not open to user queries.
Invite Members: Supports directly setting the invited member as the "Administrator" role through the invitation entry and Open API.
Data Forwarding: Supports custom Endpoints.

Unified Catalog¶

Entity Type Management: Optimized associated view configuration. On the associated view configuration page, if the current entity type has official built-in views, the system will automatically list them and enable them by default. Users can enable or disable them as needed.

April 22, 2026¶

Feature Updates¶

Unified Catalog¶

Added the "Unified Catalog" feature module: Centrally manages system services and cloud resource assets, providing entity maintenance, relationship topology analysis, and full architecture view capabilities.
Unified Asset Management: Supports 5 preset entity types: System, Service, Host, Database, Cloud Resource, and custom extensions. Supports 4 integration methods: automatic discovery, DataKit reporting, cloud asset collection, YAML declaration.
Entity Relationship Insights: The detail page displays entity metadata, contact information, and external links. Supports two topology modes: dependency view (showing call and containment relationships) and hierarchy view (showing only containment relationships). The default display shows one level of call depth, with support for switching to full path topology.
Full Architecture Topology: Automatically generates a workspace full architecture diagram, supporting multi-dimensional filtering, group display (up to 5 levels of nesting), and view saving.

Obsy AI¶

Obsy AI has been adjusted to open as a right-side drawer globally, supporting natural language Q&A, error analysis, AI-generated dashboards, etc.
When opening Obsy AI on detail pages of objects such as error logs, alert events, dashboards, error traces, and RUM Errors, the corresponding quick analysis hints are automatically displayed. One-click sending analyzes the current data.
Supports generating dashboards via natural language descriptions. After the AI generates a draft, you can modify the dashboard information and confirm the generation. After a real-time preview, it can be saved as a formal dashboard. Supports returning to the Obsy AI dialogue to redefine requirements.
AI-generated dashboards are automatically marked with an AI identifier, supporting list filtering.

Open API¶

LLM Applications: Supports full lifecycle management (create, modify, delete, list).
Unified Catalog: Supports entity and entity type management (create, edit, delete, export, list).
Supports users rotating workspace Tokens. After the operation is completed, the system automatically sends a security reminder email to the Owner and all administrators.

Incident¶

The Explorer now includes a "Create Rule" quick entry.
Supports configuring data filtering conditions when creating rules (covering all types: Logs, APM, RUM) to precisely define the effective data range. The rule list also includes a filter condition display column.

Monitoring¶

Notification Targets: For the Hong Kong and International sites, added support for creating Telegram Bot notification targets, pushing alert messages directly to Telegram private chats, groups, or channels.

Unrecovered Events¶

The event list supports dragging to customize column widths.
The df_dimension_tags field is displayed as a toggle. When enabled, it is displayed as tags in the list and supports clicking to navigate to the associated page.

Pipeline¶

Pipeline rule configuration > Filter Conditions: Supports wildcard matching (* matches any characters, ? matches a single character), allowing flexible filtering of data sources.

Before using this feature, ensure the version requirements are met: For standard Commercial Plan, DataKit must be upgraded to version 1.93.0 or higher for the local Pipeline to take effect (central Pipeline configuration is not affected by the DataKit version). For Deployment Plan, the central Pipeline must be updated to version 1.129.224 or higher.

Cross-Space Query¶

Core interfaces for APM and RUM support cross-space queries.

Scenes¶

View Variables:
- Added support for "Synthetic Monitoring" data queries, allowing selection of API Tests or Multistep Tests as the query source.
- When selecting a variable to execute a query, an "Apply" action item is added to the dropdown list. Clicking it executes the query immediately.
Dashboard > External Sharing:
- Supports locking view variables (the person shared with cannot change them).
- The time range supports selecting absolute time.

Infrastructure¶

Network Overview > Network Path > Detail Page > Network Flow: Added a "Jump to Network Flow" button, which can carry the current filter conditions to navigate to an independent network flow page.
Database:
- PostgreSQL & MySQL now include Query and Session tabs: The database detail page now includes "Query" and "Session" tabs, supporting viewing SQL query performance and session connection details (❗️Requires upgrading to the latest DataKit and enabling corresponding data collection).
- Overview page > Chart Optimization:
  - Extended attributes are now separated from the overview page into a new "Extended Attributes" tab.
  - Added a "Query Latency and Execution Count Trend Comparison Chart", showing the trend of total query count and total latency over the time range.
  - Added a "Query Wait State Distribution" pie chart, a "Load Trend" bar chart, and key metric views such as "CPU Usage", "Memory Usage", "Disk IO", and "Network IO".

Others¶

Added Personal API Key configuration.
When the browser has recorded a login state, accessing the Guance /login page automatically redirects to the homepage, preventing already logged-in users from repeatedly entering the login page.
Optimized import experience: Import conflicts are now handled by group, distinguishing three types of conflicts (ID + Name conflict, ID only conflict, Name only conflict), covering dashboards and workspace migration. The result window prompt has also been optimized.
Indexes/Global Tags: Added support for import/export.
Ticket Management: When submitting or commenting on tickets, attachments now support .mov and .mp4 formats.

Deployment Plan¶

Management Backend > Workspace List > Index Configuration: For OpenSearch/ES storage, added the rescource_map data type, supporting custom primary shard count (❗️Only affects newly generated indexes).

Bug Fixes¶

Fixed an issue where events were generated by a monitor and the user received a phone alert, but the event could not be seen in the Event Center or Unrecovered Events.
Fixed an issue with abnormal data querying for RUM in the interface.
Fixed an issue where view variables failed to reference certain numeric fields.
Fixed an issue where key-value pairs set for test tasks via tags on the Global site did not take effect even 20 minutes after configuration.
Fixed an issue where querying data for more than three hours in Infrastructure > Containers displayed as empty.
Fixed an issue where date and timestamp fields had no values when creating events via the API.
Fixed an issue where querying data via the Open API on the Global site reported an error.
Fixed an issue where the DQL query grouped table chart was rendered incorrectly.
Fixed an issue where a team could not be cleared after selection on the entity edit page.
Fixed an issue where a data source had stopped writing, but the ranking list still showed log volume growth after refreshing by source.
Fixed an issue where the data collector management page in the backend management displayed an empty IP.
Fixed an issue where the language switch button disappeared after sharing a snapshot.
Fixed an issue where the grouped bar chart should not display groups with no data for the current day in the hover tooltip.
Fixed an issue where the client reported successful data upload, but the data could not be queried on the platform.
Fixed an issue where the AI assistant on the Global site reported an error when used.
Fixed an issue where the service map feature in APM was too slow to respond and unusable.
Fixed an issue where calling the AI interface failed when opening AI analysis in the Event Center on the Global site.
Fixed an issue where the bar chart did not display expression query results and could not add aliases.
Fixed an issue where converting a simple query to a DQL query in the query tool did not convert to Base64 encoding.

April 9, 2026¶

Feature Updates¶

Monitoring¶

When creating or editing a monitor, if a field matches the high-risk field list, the system now adds detection metric configuration restrictions.

Dashboards¶

Groups: Supports configuring whether the group is expanded or collapsed.

April 1, 2026¶

Feature Updates¶

Plans & Billing¶

For the Hong Kong and Global regions, SMS and Voice Call have new pricing:

Billing Item Billed in CNY Billed in USD

SMS ¥7 $1

Voice Call ¥0.84 $0.12
Added a Data Points billing mode. To switch from "Time Series" to "Data Points" for metric billing, please contact your account manager.

Scenes¶

Chart Query > External Data Source Query: Added support for the #{step} dynamic variable:
- Supports using #{step} as an interval placeholder in query statements, replacing a fixed interval value.
- The system automatically calculates and passes the optimal interval value (in seconds) based on the current time range, dynamically optimizing data display precision with the query range.
- No need to manually adjust function parameters. The system automatically balances query performance and chart display effect for large data volume scenarios.
Dashboards:
- External database SQL queries now support variable value retrieval.
- Added support for importing Grafana dashboards.
Explorer > Sampling: A large data volume protection mechanism has been added to queries. When the query data volume exceeds a threshold, forced sampling is automatically triggered, and the sampling option cannot be disabled.

Integrations¶

DataKit Operator can now be installed on the Integrations > DataKit page.

Management¶

Data Forwarding / Pipelines: Supports LLM as a data source type.
Configuration Migration: Further supports one-click import and export of full workspace configurations, covering monitoring views, alert policies, data processing rules, role management settings, and data compliance core resources.
Role Management: When setting a member as an administrator, no billing center review is required, and the change takes effect immediately.

Monitoring¶

APM Monitor > Service Metrics: The dimension selection now includes resource field support.
Added an Alert Cache Protection mechanism:
- When the number of detected objects reaches a warning threshold, a system notification is triggered (max 1 per day). When it reaches the upper limit, the monitor is automatically suspended, and a system notification is sent.
- The storage period for unrecovered alert events has been adjusted from 30 days to 7 days.

RUM¶

View Explorer Detail Page > Fetch/XHR List: Supports displaying all network requests containing resources.

LLM Monitoring¶

The Explorer detail page has been updated with the following:
1. Added support for binding built-in views.
2. Added step-level Token consumption display.
3. Attribute Information > Search Box: Supports cross-Span synchronization of search content.

Integrations¶

Added Alibaba Cloud AnalyticDB MySQL integration.
Bug fix for the domain SSL certificate collector, fixing issues related to self-signed certificates.

Bug Fixes¶

Fixed an issue where the new version of the monitor's graded alert had defects, DQL execution failed, and events were generated by the executed data, but no data was returned by the quick query in the space.
Fixed an issue where the display was abnormal after setting role permissions twice consecutively in team management.
Fixed an issue where the left-side filter text in the Event Center was inconsistent with other locations, and the interface text had inconsistent Chinese and English.
Fixed an issue where the data volume for the last 24 hours was less than the last 12 hours in statistical data.
Fixed an issue where a large number of document count errors appeared in kodo-x logs.
Fixed an issue with abnormal display of frontend view metric values.
Fixed an issue where the data displayed for direct viewing and maximized viewing of unrecovered alert events in dashboards was inconsistent.
Fixed an issue where accessing the RUM page for a specific space reported an error.
Fixed an issue where the page displayed abnormally when querying event data across spaces in the Unrecovered Events dashboard.
Fixed an issue where the Deployment Plan base did not support adjusting the query time.
Fixed an issue where setting a default value did not take effect when the dashboard view variable enabled the "Include wildcard" option.
Fixed an issue where the email sender information was not provided for the International site so that customers could add it to the allowlist.
Fixed an issue where Session Replay images on the mobile end were loaded incorrectly, and many related requests reported errors.
Fixed an issue where the Multistep Test dashboard could not be displayed in the Chinese version of Synthetic Monitoring.
Fixed an issue where no data was found after integrating RUM, and message accumulation was discovered.
Fixed an issue where the monitor status remained "Disabled" after disabling and then re-enabling it.
Fixed an issue where the error details still displayed the "Toggle Parsing" tab even though the source map file was not uploaded for the mobile end.
Fixed an issue with abnormal team role configuration on the International site.
Fixed an issue where the parsing section of the error details page in the Error Explorer displayed abnormally.
Fixed an issue with escape characters in log queries.
Fixed an issue where the data type of behavior data fields in the space was not fixed.
Fixed an issue where downloading the DataKit installer was interrupted in the European site deployment environment.
Fixed an issue where querying host details in Infrastructure reported an error.
Fixed an issue where the log query returned incorrect results.
Fixed an issue where the "Follow Redirects" option in test tasks on the International site did not take effect.
Fixed an issue where the "Close Sampling" button could not be found in the dashboard view, and the table panel had values but the frontend showed no values.
Fixed an issue where the log storage duration in the space did not match the actual queryable log date range.
Fixed an issue where the central data processing rule did not take effect in the Deployment Plan environment.
Fixed an issue where the tags in metric management were fewer than the actual data after converting logs to metrics.
Fixed an issue where the data processing rule reported an error when parsing JSON data.
Fixed an issue where the default value setting for view variables did not take effect, and the expression query calculation result was abnormal.
Fixed an issue where the same DQL query displayed different results in the dashboard query and the quick query.
Fixed an issue where the output of the intelligent analysis feature in the Event Center on the International site was in Chinese.
Fixed an issue where the alert did not meet expectations.
Fixed an issue where log backup failed in the self-observability space.
Fixed an issue where a session had no replay button, but the view had a replay option.
Fixed a billing dimension deviation for data backup to Guance official storage, unifying the billing rule: changed from billing by compressed capacity to billing by raw uncompressed capacity, consistent with the official billing standard.

March 18, 2026¶

Feature Updates¶

DQL¶

Resource pre-aggregated data now includes ddsketch quantile data, supporting Resource-level quantile statistics. Monitors can be configured with alerts based on quantile metrics.
DQL Query: Use the percentile_from_ddsketch_payload function to query the duration_quantile field.

Example:

TRM::re('.*'):(percentile_from_ddsketch_payload(duration_quantile, 90)) {source_resource="GET /test/*anypath"} limit 1

Metrics¶

"Metrics Management" export to CSV now supports exporting tags synchronously:
- Added a "Tags" column, strongly associating the exported tags with the metric data in the current row.
- All dimension information bound to this metric is fully displayed in JSON array format.
Metric Details now includes the origin_field field, representing the original metric.

Log Explorer¶

A "Display Items" configuration entry has been added on the right side, making it easy for users to quickly adjust the display fields of the log list.

Dashboards / Monitors¶

Duplicate names are no longer allowed.
Creation and Update Restrictions
- When creating (directly or via Open API): If the name already exists → Creation is rejected and an error is returned.
- When importing/exporting/workspace migration: If a name conflict exists, you can choose to skip or overwrite.
Scope of Application
- Direct creation
- Open API creation
- Import / Export / Workspace migration

Deployment Plan Updates¶

For the Deployment Plan, administrators can modify the configuration in the backend management to specify whether MFA authentication is enabled for each SSO in multi-SSO scenarios, achieving flexible login method configuration.

March 11, 2026¶

Feature Updates¶

Infrastructure¶

Added Network Device Data Display:
- Supports automatic discovery of SNMP devices, displaying key metrics such as device status, IP, type, and interface count.
- The device detail page provides three views: Device Information, Interface Details, and Performance Metrics.
- Supports interface status filtering and monitor configuration.
- The performance metrics dashboard displays trends for CPU, memory, traffic, interface status, etc.
Added pages for Database:
- Query: SQL performance analysis, supporting slow query identification and execution sampling drill-down.
- Session: Session blocking analysis, automatically identifying blocking chains (Root Blocker/Waiter).
- Overview: Added a load trend chart and top 5 frequent queries.

Incident Center¶

Added an "Analysis Dashboard" page, providing visual statistical analysis capabilities for the full incident lifecycle.

Dashboards¶

Added support for adding tags to dashboards in batches.
Optimized sharing of long dashboard links.

APM¶

The upper right corner of the trace detail page, the resource field on the detail page > Service call relationship > Service card, now includes an associated resource topology. Clicking navigates to the topology page.
Service Details > Resource Invocation: Added a resource type toggle button (Entry Resources, All Resources).

Notification Targets¶

The notification target template now displays the corresponding name, making configuration clearer.
Added a test notification button to verify the reachability of the notification channel.

DataKit¶

Installation configuration page: Added a "LOG" configuration module. Supports adding file path collection log configurations, with options for auto-multiline mode and enabling native direct write index.

Management¶

API Keys Management:
- Supports editing the API Key name and role.
- The Open API address is automatically displayed at the top of the list based on the workspace's site.
Ticket Management:
- Optimized ticket types, adding type descriptions.
- Optimized the ticket search function, supporting search by user email and submitter.
- Added priority selection for tickets.
- Added support for deleting tickets and deleting comments within tickets.

Time Widget¶

Added "Recently Used Time Ranges" record: The last 5 time range query conditions used locally.

Integrations¶

Added Volcengine VCM Kafka integration.

Bug Fixes¶

Fixed an issue with an excessive number of specific write 429 errors in kodo-x logs.
Fixed an issue where querying metering data on the PaaS platform returned inconsistent results each time.
Fixed an issue with dashboard display optimization.
Fixed an issue where clicking to add a field to the filter value did not work after opening log details in a custom log Explorer.
Fixed an issue where test data could not be queried after upgrading the base to a specific version.
Fixed an issue where RUM data had values in a new space, but the Explorer query was empty.
Fixed an issue where aggregate functions were not displayed correctly in metric queries within the space.
Fixed an issue with spelling and grammar errors in the warning message on the invitation page for the overseas site.
Fixed an issue where the left-side filter button in the Log Explorer failed to correctly display the log source.
Fixed an issue where the Incident Center could not receive incident notifications via test email.
Fixed an issue with the quick query function.
Fixed an issue where the intelligent analysis feature in the Guangzhou site reported an error when used.
Fixed an issue where a 414 error was returned when the URL was too long.
Fixed an issue where parameters added to a dashboard link could not be saved if they were too long.
Fixed an issue with non-standard prompt text in the share name input box when creating a share.
Fixed an issue where querying host data in the Resource Catalog for more than one day returned no results.
Fixed an issue where a voucher was not fully consumed, but an overdue amount was generated by the Incident center, causing a false overdue warning for the space.
Fixed an issue where the APM menu on the overseas site shifted left and right during use.
Fixed an issue where the query tool on the Metrics Analysis page returned results limited to 2000, not reaching the expected maximum of 10,000.
Fixed an issue where importing alert rules across spaces reported an error.
Fixed an issue where the Service Map did not display call relationships.
Fixed an issue where the page reported an error when modifying external index rules.
Fixed an issue where loading the log detail page reported an error when opening an external log index.
Fixed an issue where the dashboard snapshot link sent via scheduled report had no data on the page after login and redirect.
Fixed an issue where the recent X hours aggregation in the dashboard could not hit the streaming aggregation.
Fixed an issue where profiling data collected via PHP APM on the overseas site could not be displayed correctly.

February 4, 2026¶

Feature Updates¶

Incident Center¶

The original "Issue" feature has been fully upgraded to the "Incident Center".

The Incident Center provides integrated incident handling support. When a monitor detects an anomaly, it automatically generates an incident event, merges duplicate alerts, and notifies the responsible person according to on-call rules. If not handled in time, the escalation strategy expands the notification scope. On the incident detail page, you can view related monitoring metrics, error logs, call traces, etc., in one place, supporting status transitions and team collaboration, with all operations fully recorded. The Incident Center will further help teams standardize incident handling processes and improve response efficiency and process transparency.

In the billing logic of the Incident Center: Each time an escalation strategy is triggered, 100 Triggers are recorded when the notification is sent.

Incident¶

The "Incident" feature is now live! It automatically aggregates errors from APM, RUM, and Logs, and intelligently converges the same issues into a unified Issue for tracking. Before use, you need to configure delivery rules to set the monitoring scope. You can then view error summaries, processing statuses, and occurrence trends in the list, or enter the detail page to analyze the full stack, related traces, and user sessions. All errors support status transitions and team collaboration, enabling full lifecycle management from discovery to resolution.
Added "Error Record Count" billing, which counts the number of new Issue data records per day, including Issue data generated by the Incident center.

Open API¶

Resource Catalog: Added support for creating, editing, and deleting resource group information.
Supports directly editing account status (On-Call, On Leave).

Metrics Analysis¶

Added Top N Series and Max Return Points options, allowing you to specify the number of data series returned (20/50/100/500) after sorting by largest or smallest.
Added support for clicking data points on the chart, with a dropdown to view similar trend metrics, drill-down analysis, or other related views.

Infrastructure¶

Hosts:
- Added support for list filtering through the df_mute field.
- For hosts globally muted via Open API or rules, the system will support displaying a "Muted" indicator in the host list.
Resource Catalog: Added a "Service List" entry.

Scenes¶

Dashboards: Added a Related Monitors button, supporting one-click viewing of monitors associated with the dashboard.
Charts: Added a unified sequence identifier for all chart alias configurations and hover-linked visualization to display the correspondence between multiple query line configurations more intuitively.

APM¶

Profiling: If the Profile file size exceeds 20MB, the system does not support online parsing. A friendly prompt has been added, suggesting the use of professional analysis tools for viewing.

LLM Monitoring¶

In the LLM Explorer "All Traces" list, "Total Tokens" now counts the total Tokens consumed by the entire Trace. The Total Tokens column will also display the Input and Output Token counts.

Logs¶

Explorer: After selecting "Reset to Default Fields" in Display Items, the message field display logic has been optimized.

Management¶

SSO Management: Optimized the SSO login flow. Users must first select an identity provider via email and complete authentication. Only after successful authentication can they view accessible workspaces in a protected state, preventing the leakage of permission information.

Deployment Plan¶

Management Backend > Global Configuration: Added Platform-level System Announcement Management Configuration.

Integrations¶

Added RedPeaks SAP integration.
Updated AWS RDS MySQL dashboard.
Added KingbaseES monitor.
Updated English dashboards, mainly handling Chinese-English conversion issues.
Updated Tencent PGSQL dashboards & monitors.
Updated Resource Catalog icons and categories.

Bug Fixes¶

Fixed an issue where dashboards containing external data sources could not load data after cloning or exporting and re-importing.
Fixed an issue where inner-dataway panicked after upgrading the platform base.
Fixed an issue with MCP Server service call support on the site.
Fixed an issue where running Func scripts in the Deployment Plan timed out, and both DataWay and Kodo reported errors.
Fixed an issue where the hosts and pods associated with the infrastructure in the service topology had no data.
Fixed an issue where the Ningxia site could not receive phone alert notifications.
Fixed an issue where the configuration information was lost after saving and reopening the monitor configuration.
Fixed an issue where trace data was not refreshed after switching services in the APM feature.
Fixed a compatibility issue where PromQL queries reported errors through the Grafana plugin.
Fixed a compatibility issue with dashboard variables using database SQL queries.
Fixed an issue where the backend displayed the number of DKs as 0 after upgrading the base.
Fixed an issue where the number of services in Metric Explorer and Monitor did not match.
Fixed an issue where setting multiple arbitrary tags in the data access metric data access feature caused the metric function to be unavailable.
Fixed an issue where the log monitor could not alert, and the execution log showed a 503 error.
Fixed an issue where the Explorer reported a query failed error when viewing an external log index.
Fixed an issue where saving an LLM application reported a 404 error.
Fixed an issue where adding an expression query to a log detection monitor reported an error.
Fixed an issue where the Deployment Plan Owner had no permission to save security check rules.
Fixed an issue where the Grafana plugin reported an error when there were many query statement interfaces.
Fixed an issue where creating a security check reported an error.
Fixed an issue where the monitor execution timed out.
Fixed an issue where the log quick filter items were returned by the backend but the field value list was not displayed.
Fixed an issue with URL handling in custom event notification templates.
Fixed an issue with viewing and modifying dashboard groups.
Fixed an issue where the geoip() function in Pipeline parsed public IPs in logs as all nulls and was ineffective, and the parsed fields were not displayed on the test UI page.
Fixed an issue where costs were still incurred 7 days after a workspace was dissolved.
Fixed a bug with the K8s built-in view.
Fixed an issue where inviting users in the European space reported a timeout error.
Fixed an issue where the DQL statement {condition1 OR condition2} only returned results matching the first condition.
Fixed an issue where submitting an issue ticket in the workspace ticket system reported an error.
Fixed an issue where the popup in Metric Analysis, when no metric was selected, displayed both metrics and tags.
Fixed an issue where the alert value for log alerts did not match the actual query value.
Fixed an issue where clicking to jump to Pod monitoring from Infrastructure Containers inexplicably redirected to the login page.

January 22, 2026¶

Feature Updates¶

Logs¶

Log index names now support the hyphen - naming format.

Management¶

Data Forwarding: When configuring forwarding rules with logs as the data source, you can now select native direct write index filter conditions.
Member Management: Added a "Show only enabled permissions" toggle.
Preferences: Added a default entry page option for APM, allowing you to customize the default page (Traces, Services, Analysis Dashboard) when clicking the APM navigation bar on the left.

Time Widget¶

Time Widget Enhancement: When the system detects that the local device time differs from the server time by more than 1 minute, it will automatically display a deviation prompt, helping to avoid data query range errors caused by inaccurate local time.

January 14, 2026¶

Feature Updates¶

Monitoring¶

For overseas sites, added Google Chat as a notification target.

APM¶

Service Details:
- Resource Invocation > Upstream/Downstream Invocation: Enhanced interaction effects:
  - Quick expansion: When hovering over a resource node, a "+" button appears. Clicking it quickly loads its upstream and downstream. The start and end nodes display "Start"/"End" identifiers.
  - Connection line display metrics: The connection lines directly display request count, P99 latency, and error rate, accompanied by a legend.
  - Added loading prompt: A "Loading" state is displayed when expanding the topology.
  - One-click fullscreen: Added a fullscreen button, supporting one-click enlargement/restoration of the topology.
- Service Map: Added an "Infrastructure" option. Clicking it supports switching to display the "Host" or "Pod" topology.
- Added an "Infrastructure" tab: Supports switching between Host and Pod views, displaying host trend charts and lists. Clicking a list item expands the detail page for more information.
Traces > Detail Page: Associated Logs: Optimized the index filter function. By default, all indexes of all associated services within the current trace are selected.

Scenes¶

Dashboards:
- Supports sharing dashboards.
- Added support for tag permission management page, controlling dashboard viewing and management permissions uniformly through tags.
Chart Query: When selecting in and not in filter conditions, you can now use the result of a subquery as the filter value.
Chart Sharing: Added support for sharing charts via link.

Logs¶

Added index name search functionality, automatically counting the number of each index type returned by the search, and positioning to the tab of the index type with the highest count.

Management¶

Share Management:
- Shared snapshots/dashboards: Added a language switch button to meet the needs of global users.
- Shared Charts: Added embed code/link toggle buttons, allowing switching between displaying the corresponding chart.
- Added Shared Dashboards management page.
API Key Management: Expanded the permissions of API keys to meet the advanced operational needs of the Owner role.
Role Management: Added a "Show only enabled permissions" toggle.

RUM¶

Expanded the SDK integration configuration for Android, iOS, and React Native, adding advanced feature configuration items for RUM, Traces, and Logs, and highlighting them when modified.

Plans & Billing¶

Optimized the aggregation logic for monthly bills, adding daily bills (from the 1st to the query date) for all cumulative consumption.

Integrations¶

Added OBCloud OceanBase integration documentation.
Optimized zabbix-api/stream integration documentation.
Optimized Go integration documentation, adding 1.18-2.23 integration methods for compile-time instrumentation.
Added Resource Catalog templates: tencentcloud_clb, aliyun_cnapigateway, aliyun_slb, aws_api_gateway.

Bug Fixes¶

Fixed an issue where the corresponding service was not displayed in the Log Explorer's quick filter.
Fixed an issue with abnormal display of quick filters.
Fixed an issue where the system reported an error after upgrading to a specific version.
Fixed an issue where clicking the sorting function in the grouping mode of Metrics Analysis did not respond.
Fixed an issue with unlimited batch operations on detection rules.
Fixed an issue where P90 data could not be displayed in dashboards.
Fixed an issue with incorrect display of Span parent-child relationships in the APM waterfall chart.
Fixed an issue where the source configuration in the log Pipeline edit view on the overseas site could not be filled correctly.
Fixed an issue where using the int function in DQL reported an internal server error.
Fixed an issue where the space authorization metric supported interface filtering, but the OpenAPI did not provide the corresponding function.
Fixed an issue with inaccurate log query results on the Oregon site.
Fixed an issue where previously working data access rules on the overseas site suddenly failed, and users with different roles reported errors when accessing logs.
Fixed an issue where clicking the APM > Services menu reported an error.
Fixed an issue where changing the query time window in a dashboard on the overseas site caused the chart to report an error.
Fixed an issue where the link "Go to Workspace for AI Analysis" in alert notifications was incorrect.
Fixed an issue where the "Export CSV" button disappeared from the Log Explorer, and logs for a specific date could not be viewed.
Fixed an issue where the monitor name in the Event Center did not display a value, but the backend interface returned data.
Fixed an issue where the certificate was not set as optional when selecting SASL_SSL protocol in Kafka data forwarding configuration.
Fixed an issue with abnormal display of the status chart on the APM monitor page.
Fixed an issue where the monitor alert could not be sent and showed a 429 error.
Fixed an issue where the dropdown input box scrolled with the page.
Fixed an issue where the data trend chart could not be displayed normally on the configure monitor page.
Fixed an issue where selecting "Last 12 hours" on the log performance interface caused the left-side "Status" and "Service" filters to take more than 2 seconds to load.
Fixed performance issues where querying logs for specific services/conditions, waiting for query results, and loading single log details had excessively long response times (below standard, over 8 seconds, and over 0.3 seconds, respectively).
Fixed an issue where the default alert threshold for Linux host disk usage needed adjustment.
Fixed an issue with viewing Span count distribution when querying specific services and resources and switching to "Time Series" chart mode.
Fixed an issue where changes to notification targets or notification targets in alert policies did not take effect.
Fixed an issue where an error was reported during initialization in a specific version environment.
Fixed an issue where view variable filtering did not work in the Infrastructure > Containers > Analysis Dashboard on the overseas site.
Fixed an issue where the interface displayed an internal server error when aggregating RUM data on the "Metrics > Metrics Analysis" menu on the overseas site.
Fixed an issue where selecting a long log line in the Log Explorer list mode did not automatically wrap, and the stacked mode also had no default wrapping.
Fixed an issue where the time sort in the Log Explorer did not take effect, causing time display confusion.
Fixed an issue with inaccurate query values for the number of Triggers in metering data.
Fixed an issue where profiling data collected via PHP APM on the overseas site could not be displayed correctly.
Fixed an issue where clicking on profiling data to view details after enabling the profiling feature of the Node.js data collector on the overseas site reported an error.
Fixed an issue where after selecting a default variable in the dashboard view variable three-level linkage, the other three view variables were displayed as empty, requiring a default data item to be displayed.
Fixed an issue where the data access rules set for RUM on the overseas site did not take effect.
Fixed an issue where the rules configured for the user access scope in the data access menu on the overseas site caused matching errors.
Fixed an issue where all statistical data was abnormal when testing the funnel chart feature and adding a count item on the overseas site.
Fixed an issue with the incorrect field type for query_range in audit logs.

January 6, 2026¶

Feature Updates¶

Scenes¶

Added Data List: Compatible with the original "Log Stream Chart" and "Object List". These two chart types will be deprecated in the future.
- The system pre-configures common default display columns for each data source. You can directly select the fields to display in the data list above.
- Common Configuration: Added "Alias" configuration. When the data source is Logs, added a "Display" configuration.

APM¶

Fixed an issue where the downstream call attribution of the middleware consumer end was incorrect in the service call relationship diagram, restoring the true service call relationship of a single trace.

Billing Item	Billed in CNY	Billed in USD
SMS	¥7	$1
Voice Call	¥0.84	$0.12