Issue:
System Pipeline for the job has not stopped after stopping the job.
Solution:
In the Control Hub (SCH) version 3.2.0.0, we introduced an enhancement - Graceful termination of system job. Previously, the System Pipeline was abruptly stopped. This might have resulted in some of the metric records in Stats Aggregator stage to be not processed.
How is it handled now?
When an actual job is stopped, the Control Hub will wait for System Pipeline to signal that the pipeline has finished sending all metric records to it. Once the Control Hub gets this signal, it triggers a stop of System Pipeline. In a worse case where System Pipeline can’t send this signal (for example, due to an error), the Control Hub will wait for 'Pipeline Force Stop Timeout' before forcefully shutting down system pipeline. By default, the Pipeline Force Stop Timeout is set to 2 minutes and configurable through the job UI.
The user can find the status of System Pipeline in the job view.
