When they run out of disk space or memory they may stop showing logs or even stop entirely prematurely.
When some of those timeouts happen, the workflows in gameci/docker repo will not report back
In these latest changes we're reporting everything that versioning backend does to #backend channel on discord and recover from builds that have status "started" but are in fact already failed or succeeded.
Previous unreviews PRs
These two are part of the changes, but I merged them to check and update the deploy flow.
So they're still fine to be reviewed as part of code review.
If an image never reported back after 6 hours, we mark the build as either failed or published, based on whether the docker image has the "most specific tag" assigned to it (which includes baseOs and the most specific repositoryVersion.
Other notes
I've already put these changes live from my local, so everything works.
Checklist
[x] Read the contribution guide and accept the code of conduct
…erhub
Context
#backendchannel on discord and recover from builds that have status "started" but are in fact already failed or succeeded.Previous unreviews PRs
These two are part of the changes, but I merged them to check and update the deploy flow. So they're still fine to be reviewed as part of code review.
Changes in this PR
baseOsand the most specificrepositoryVersion.Other notes
Checklist