Dedicated Compute Node Service¶

The Dedicated Compute Node Service is a Guance exclusive computing power solution designed for high-performance and high-stability scenarios. Compared to the default shared compute pool model, dedicated compute nodes provide physically isolated computing resources for a single tenant to handle query and analysis tasks for logs, metrics, traces, and other data, meeting the observability requirements of production-level core business scenarios.

Core Mechanism Differences¶

In the default version, Guance employs a shared compute pool architecture. Query requests from all tenants are scheduled through a unified public Query Queue, with tenant-level concurrency control ensuring overall system stability. This means:

• Tenant query concurrency has a platform-level upper limit, potentially causing queuing during peak periods.
• Concurrency for large-scale or complex query tasks (e.g., aggregation across long time ranges) is restricted.
• High-frequency OpenAPI calls may trigger 429 rate limiting.

Dedicated compute nodes completely resolve the above limitations through a tenant-level resource isolation mechanism:

Applicable Scenarios¶

Dedicated compute nodes are particularly suitable for scenarios requiring high query stability and concurrency capabilities:

1. High-Frequency OpenAPI Calls

High-frequency API call scenarios such as programmatic automated queries, scheduled data retrieval tasks, and CI/CD integrated analysis, avoiding concurrency limits of the shared pool.

2. Multi-Team Parallel Analysis

Multiple teams simultaneously performing data exploration and troubleshooting through the console, requiring high query concurrency support.

3. SRE / Operations Automation

Agent-triggered automated queries, anomaly analysis, and decision execution (e.g., auto-scaling judgments).

4. Security and Compliance Analysis

Large-scale log search and correlation analysis, involving multi-dimensional aggregation calculations across terabytes of data.

5. FinOps Cost Analysis

Cross-dimensional, multi-period data aggregation calculations requiring stable computational performance guarantees.

6. AI / Agent-Driven Systems

Using observability data as real-time input for AI or business systems, demanding millisecond-level stable response.

Core Capabilities¶

1. Dedicated Computing Power Isolation

Allocates an independent computing resource group (Query Group) for a single tenant, avoiding resource competition with other tenants and ensuring stable performance.

2. Higher Query Concurrency

Supports significantly increased query concurrency to meet the needs of multiple users and systems accessing simultaneously, with specific upper limits calculated independently based on node configuration.

3. Enhanced Large Query Capability

Optimized for queries with large time ranges, high data volumes, and complex aggregations, significantly improving execution efficiency and reducing timeout risks.

4. Stable Query Performance

Maintains stable response times even under high-load scenarios, eliminating performance jitter caused by overall load fluctuations in the shared pool.

5. Support Capability for AI / Agent

Provides reliable data query infrastructure for automated analysis, intelligent operations (SRE Agent), security analysis, FinOps, and other scenarios, ensuring API call success rates and response latency.

Deployment and Scaling¶

Dedicated compute nodes support on-demand configuration and smooth scaling:

• Supports configuring computing resources (CPU / Memory / Concurrency) of different specifications based on business scale.
• Supports subsequent on-demand scaling (horizontally increasing node count or vertically upgrading single-node specifications).
• Seamlessly integrates with the existing Guance data system, requiring no modifications to data collection configurations.

Activation Method¶

The Dedicated Compute Node Service requires separate application for activation and is billed based on resource configuration. If you wish to improve query performance or evaluate adaptation solutions, please contact your account manager. We will provide customized configuration recommendations based on your business scale and usage scenarios.