Task hub worker stops and restarts constantly. Is this expected?

[jbarntish]

Hi all,

According to our app insights logs, it looks like the task hub worker is stopping and starting over and over when no orchestrations are running. Is this normal/expected behavior? Looked through other tickets, posts and such and didn't find anything specific.

I'm just more or less looking to gut check that this is normal operation and not something to be concerned about.

Thank you.

[sebastianburckhardt]

Is this a consumption plan? On consumption plan, the task hub can scale to zero (the worker stops when there is no activity for a while but gets restarted if there is activity again).

[jbarntish]

This is an EP2, with 1 node and scale-out set to 5 nodes.

[sebastianburckhardt]

With nothing running, on an EP2, the expected behavior is that the scaler keeps one host running continuously, without frequent restarts. So what you are seeing would be anomalous.

From the traces it is unclear what is causing the host to restart; it may easily be caused by components other than durable functions or Netherite. For example, functions that time out cause host restarts. Do you have any other functions running in the app?