Worker stops polling the temporal server cluster

[Arun-V00222]

We are using temporal version 0.2.0 in production and are observing the below issue

Container - ECS
Instance - EC2

On startup the temporal client is able to poll for activities & workflows available to process and after sometime the polling stops. But the process is still running.

So everytime the process has to be restarted for client to start polling again.

[hughevans]

Are you able to inspect the process? We’re experiencing a similar issue and found that the whole Ruby process for the worker is in a very odd state, not even responding to kill -9.

[Arun-V00222]

but we are able to kill the process by logging into the container. Supervisor automatically restarts the process and the polling works fine then.

[DeRauk]

Thanks for reporting the issue. When you kill the process is there a stack trace printed out?