Managing Jobs

Every performance test and density test creates a job — a tracked unit of work with a unique ID, state, timing information, and results. The Jobs page lets you monitor running tests, review completed results, and stop jobs that are no longer needed.

Job states

Each job transitions through the following states:

State	Description
RUNNING	The job is currently executing. Pipelines are active and metrics are being collected.
COMPLETED	The job finished successfully. Full results (FPS, output paths) are available.
FAILED	The job terminated with an error, or was cancelled before producing valid results.

Job information

The Jobs page displays the following information for each job:

Field	Description
Job ID	Unique identifier assigned when the job is created
Type	performance or density
State	Current state (RUNNING, COMPLETED, or FAILED)
Start time	When the job was created
Elapsed time	Duration from start to current time (or end time if completed)
Details	Human-readable status messages (e.g., current iteration in density testing)

Viewing job results

Click on a completed job to view its detailed results:

• Performance jobs — Total FPS, Per Stream FPS, stream count, output video paths, and live stream URLs.

• Density jobs — Maximum sustainable stream count, per-stream FPS at that count, and stream distribution across pipelines.

• Latency metrics — If latency metrics were enabled for the job, the results include avg/min/max latency values (in milliseconds) per reporting interval.

Stopping a running job

To stop a job that is currently running:

1. Locate the job in the list (state = RUNNING).

2. Click the Stop button.

3. The job will be terminated.

Behavior after stopping:

• Performance tests — If the pipeline had already started processing frames, partial results (FPS, output videos) may be preserved and the job state transitions to COMPLETED. If no frames were processed, the job transitions to FAILED.

• Density tests — Cancelled density tests always transition to FAILED because partial results from an incomplete search algorithm are not meaningful.

Note: Stopping a job sends a cancellation signal to the underlying GStreamer subprocesses. The actual termination may take a few seconds while pipelines flush their buffers.

API endpoints

Jobs can also be managed programmatically via the REST API:

Operation	Endpoint	Method
List all performance jobs	/api/v1/jobs/tests/performance/status	GET
Get performance job status	/api/v1/jobs/tests/performance/{job_id}/status	GET
Stop performance job	/api/v1/jobs/tests/performance/{job_id}	DELETE
List all density jobs	/api/v1/jobs/tests/density/status	GET
Get density job status	/api/v1/jobs/tests/density/{job_id}/status	GET
Stop density job	/api/v1/jobs/tests/density/{job_id}	DELETE

For full API documentation, see the auto-generated OpenAPI docs at /docs on the backend service.