Performance of sentry worker

1

I am running Senty in private Kubernetes cluster using Sentry official image.

I set confinguration like below for HA.

web:
  replicas: 8
  resources:
    limits:
      cpu: 1
      memory: 1Gi
    requests:
      cpu: 250m
      memory: 250Mi

worker:
  replicas: 8
  resources:
    limits:
      cpu: 2
      memory: 2Gi
    requests:
      cpu: 250m
      memory: 250Mi

cron:
  replicas: 1
  resources:
    limits:
      cpu: 2
      memory: 2Gi
    requests:
      cpu: 250m
      memory: 250Mi

But worker pod’s memory and cpu usage is too high.

Memory usage was low when i first created, but it is over 95%.

And, worker’s cpu usage somtimes is over 90% even though the request volume has not increased.

workers show these error messages. (Could it be the cause?)

06:33:33 [ERROR] multiprocessing: Process 'Worker-130' pid:3055 exited with 'signal 9 (SIGKILL)'
08:11:06 [ERROR] multiprocessing: Process 'Worker-144' pid:3083 exited with 'signal 9 (SIGKILL)'
09:28:58 [ERROR] multiprocessing: Process 'Worker-111' pid:3017 exited with 'signal 9 (SIGKILL)'
11:17:58 [ERROR] multiprocessing: Process 'Worker-127' pid:3049 exited with 'signal 9 (SIGKILL)'

So, my question is,

is it normal to run sentry with so much cpu & memory usage, or maybe should i increase replicas or resources?
Or, Could you give me any other advice?

1 Like
2

Sorry to push but this is still a problemt that persists! Any advice is appreciated!

1 Like
3

We are seeing the same, workers are consuming a lot of resources - are there any recommendations on how to set up deployments for optimal resource utilization?