We're really grateful for Windows Runners since it's solved our on premise deployment issue. Also it's allowed us to finally build / deploy some older .NET Framework projects that aren't dockerized.
My only issue is that it seems the runner isn't too reliable. Or it could be our CI/CD server - I'm just not sure how to tell. Issues I see randomly through the day
- some star network error in the pipelines saying it can't communicate to the runner
- the runners just super slow to start up
- new pipelines waiting in Queue for a long time when nothing else is building
- sometimes there will be super large time gaps in between steps - but the step still says it's fast
- i.e step takes 10-15 minutes to complete an "echo" but once it completes it says 3s
I just restarted the CI/CD server and it takes several minutes to do it's first heartbeat
[2022-04-29 18:01:27,936] Runner version: 1.313
[2022-04-29 18:01:28,195] Runner runtime: windows-powershell
[2022-04-29 18:03:33,624] Starting websocket listening to RUNNER_UPDATED events.
[2022-04-29 18:03:56,263] Updating runner status to "ONLINE" and checking for new steps assigned to the runner after 0 seconds and then every 30 seconds.
[2022-04-29 18:04:22,110] Updating runner state to "ONLINE".
[2022-04-29 18:04:44,136] Updating runner state to "ONLINE".
[2022-04-29 18:04:45,715] An error occurred whilst updating runner state to "ONLINE".
com.atlassian.pipelines.stargate.client.core.exceptions.StargateConflictException: Response Summary: HttpResponseSummary{httpStatusCode=409, httpStatusMessage=Conflict, bodyAsString={"key":"agent-service.runner.conflict","message":"Simultaneous state updates were attempted for runner with id: {1ff233e2-333a-5921-a31b-00b7ac6cfbf3}","arguments":{}}}
I just deployed the same project twice and it took 28m the second time vs 1m the time before. here's a project using runner's durations for a few deploys - about 1-2m is what I'm expecting
- 28m 36s
- 1m 32s
- 3m 30s
- 1m 42s
- 7m 3s
