This message was deleted.

So, here's the message where we can tell the container is picking it up:

{"caller":"vcs/github_client.go:497", "<http://cos.googleapis.com/container_id|cos.googleapis.com/container_id>":"8b5286e31da4561f8692ae423419d7e3d8d9b430c092a1cc47422fe712a95bf1", "<http://cos.googleapis.com/container_name|cos.googleapis.com/container_name>":"klt--ejnq", "<http://cos.googleapis.com/stream|cos.googleapis.com/stream>":"stdout", "json":{…}, "level":"debug", "msg":"AllowMergeableBypassApply feature flag is enabled - attempting to bypass apply from mergeable requirements", "ts":"2023-07-26T16:28:58.318Z"}

maybe there is a permission missing for the atlantis user or github action?

I don't think so. I mean, atlantis works if we turn off the thing in Github requiring that status check.

it is possible that an error like that was not handled and does not show for some reason

And then here is where the command failed:

{"caller":"events/instrumented_project_command_runner.go:83", "<http://cos.googleapis.com/container_id|cos.googleapis.com/container_id>":"8b5286e31da4561f8692ae423419d7e3d8d9b430c092a1cc47422fe712a95bf1", "<http://cos.googleapis.com/container_name|cos.googleapis.com/container_name>":"klt--ejnq", "<http://cos.googleapis.com/stream|cos.googleapis.com/stream>":"stdout", "json":{…}, "level":"error", "msg":"Failure running apply operation: Pull request must be mergeable before running apply.", "stacktrace":"<http://github.com/runatlantis/atlantis/server/events.RunAndEmitStats|github.com/runatlantis/atlantis/server/events.RunAndEmitStats>
	<http://github.com/runatlantis/atlantis/server/events/instrumented_project_command_runner.go:83|github.com/runatlantis/atlantis/server/events/instrumented_project_command_runner.go:83>
<http://github.com/runatlantis/atlantis/server/events.(*InstrumentedProjectCommandRunner).Apply|github.com/runatlantis/atlantis/server/events.(*InstrumentedProjectCommandRunner).Apply>
	<http://github.com/runatlantis/atlantis/server/events/instrumented_project_command_runner.go:46|github.com/runatlantis/atlantis/server/events/instrumented_project_command_runner.go:46>
<http://github.com/runatlantis/atlantis/server/events.runProjectCmds|github.com/runatlantis/atlantis/server/events.runProjectCmds>
	<http://github.com/runatlantis/atlantis/server/events/project_command_pool_executor.go:48|github.com/runatlantis/atlantis/server/events/project_command_pool_executor.go:48>
<http://github.com/runatlantis/atlantis/server/events.(*ApplyCommandRunner).Run|github.com/runatlantis/atlantis/server/events.(*ApplyCommandRunner).Run>
	<http://github.com/runatlantis/atlantis/server/events/apply_command_runner.go:166|github.com/runatlantis/atlantis/server/events/apply_command_runner.go:166>
<http://github.com/runatlantis/atlantis/server/events.(*DefaultCommandRunner).RunCommentCommand|github.com/runatlantis/atlantis/server/events.(*DefaultCommandRunner).RunCommentCommand>
	<http://github.com/runatlantis/atlantis/server/events/command_runner.go:296|github.com/runatlantis/atlantis/server/events/command_runner.go:296>", "ts":"2023-07-26T16:30:36.979Z"}

well if plan status worked this should

Yeah.

yes

I'm not seeing that in the logs. 😕

that is why I mentioned, maybe we need more debugging there

I'm not sure why the ones that should be there aren't there though. That's what's confusing to me 😕

if you have the time please do not hesitate to contribute

Yeah, I would, I'm just not sure why it's not working. My Go is mid at best. I'm going to keep poking at it this afternoon though to see if I can figure out what's going on.

pre_workflow_hooks is a separate stack call

Shouldn't it be using the same logger though?

I think the instrumentated runner is eating it

ohhh yes, you were looking at the pre-workflow-hook code which is not the same

Ah...okay. I'm just not sure how to get some more insight there into what's going on. 😕

honestly, I'm not familiar with the instrumentation/metrics code just yet as we don't have it setup at Autodesk

I'm mostly wondering if anyone else who's tried to use

ATLANTIS_GH_ALLOW_MERGEABLE_BYPASS_APPLY

has gotten it working properly, and if they ran into any trouble.

when you said you have a different name on the check I wonder about that, so one atlantis is picking up the event and not able to update it

That is what I'm suspicious of. We have an

atlantis-nonprod/apply

and an

atlantis-prod/apply

status check, both are required green to merge. My assumption would be that if nonprod didn't see any changes though, it would automatically be green, but I wonder if it momentarily goes into a wait state because it picks up the atlantis command.