wiki/Pipeline-server.md
2024-03-19 10:19:04 +01:00

1.9 KiB

Pipeline server debugging

Logs from the pipeline server are now available on kibana.

Example query strings are:

  • "message": “name:<PIPELINE_NAME>" (for pipeline logs)
  • "message": "camera:<CAMERA_NAME>" (for camera logs)
  • "host.hostname":"sf-daqsync-XX.psi.ch" (for server logs)
  • “systemd.unit":”XXX.service" (for service logs)

https://kibana.psi.ch/s/camserver/app/logs

To check the metrics of the cluster (CPU. Memory, network):
https://kibana.psi.ch/s/camserver/app/metrics/inventory

Or a specific server directly:
https://kibana.psi.ch/s/camserver/app/metrics/detail/host/sf-daqsync-03.psi.ch

To monitor service availability and automatic alarm reporting (by email):
https://kibana.psi.ch/s/camserver/app/uptime https://kibana.psi.ch/s/camserver/app/observability/alerts?workflowStatus=open

Dashboards to display individual pipeline metrics. E.g this one for PSSS:
https://kibana.psi.ch/s/camserver/app/dashboards#/view/058f4b00-87a3-11ee-8a7f-6175eba7c3d6?_g=(filters:!(),refreshInterval:(pause:!t,value:0),time:(from:now-1h,to:now))

Dashboards to stream individual pipeline metrics. E.g this one for SAROP21-PBPS103_proc: https://kibana.psi.ch/s/camserver/app/logs/stream?flyoutOptions=(flyoutId:!n,flyoutVisibility:hidden,surroundingLogsId:!n)&logPosition=(end:now,position:!n,start:now-1d,streamLive:!t)&logFilter=(language:kuery,query:%27%22message%22:%20%22name:SAROP21-PBPS103_proc%22%27)

Note the stream may take around 30 seconds to load

Cam Server git

https://github.com/paulscherrerinstitute/cam_server/blob/master/cam_server/pipeline/data_processing/functions.py