Data Observability¶

Guance Data Observability helps you understand the running status of data processing jobs and the sources of data assets and their upstream/downstream dependencies. You can view job execution results, trace data sources, and, when a job fails, troubleshoot downstream impact along the lineage.

For example, an order raw table passes through a cleaning job to generate an order detail table, which is then read by a summary job. Data Observability displays these collected read/write relationships as:

flowchart LR
    A[Order Raw Table] --> B[Order Cleaning Job]
    B --> C[Order Detail Table]
    C --> D[Order Summary Job]
    D --> E[Order Summary Table]

Core Features¶

In the console, open Data Observability and the Overview page is displayed by default. You can switch modules based on your troubleshooting target:

Module	Purpose
Overview	View run counts, failed runs, failed jobs, success rate, and run trends; start job troubleshooting from the most recent failed runs.
Data Catalog	Search data assets and view associated runs, field structure, direct relationships, and extended attributes.
Jobs	View the job list and run trends; open a single run to check status, duration, inputs/outputs, and error messages.
Data Lineage	View upstream/downstream relationships around one or more assets to trace origins or analyze downstream impact.

Core Objects¶

Object	Description
Dataset	A data asset read or produced by a job, such as a table, view, Topic, Stream, or file dataset.
Job	A data processing task that can be executed repeatedly, such as an order cleaning job.
Run	A single execution of a job. You can view the status, duration, and inputs/outputs of each run.
Lineage	The dependencies between data assets and jobs, describing where data comes from, which jobs it passes through, and where it flows to.

Get Started¶

After data ingestion is complete, you can view data as follows:

Go to Data Observability > Overview to view run results.
In Jobs, select the job to troubleshoot, enter Run History, and view the status, error messages, and inputs/outputs of a single run.
In Data Catalog, open the relevant data asset to view its field structure and associated jobs.
Click View Lineage or View Downstream Impact to continue troubleshooting along the upstream/downstream relationships.

Viewing data requires the data observability query permission of the target workspace.

View Data¶

Data Observability displays collected information. When a field shows "Not reported", "Unknown", or "—", the corresponding information is not yet available.

Storage & Billing¶

Data Observability supports multiple data retention durations. For configuration, see Data Storage Policy.

The Commercial Plan is billed by valid Run count, while the Enterprise Plan is billed by write traffic and storage volume. For details, see Commercial Plan Billing Logic and Enterprise Plan Billing.Free Plan quota: see Plan Differences.