Deploys sometimes fail silently when automerge-action times out

[roryabraham]

Related issue: pascalgn/automerge-action#122

Problem

In workflow runs such as this one, a production deploy silently failed to run because the https://github.com/pascalgn/automerge-action silently failed.

Why this is important

We need to be notified when this action fails so we can be made aware of the failed deploy and address it if necessary.

Solution

Fix the issue in the linked repo, then update this repo to use the patched version.

[roryabraham]

We later discovered that the automerge action was likely failing because we had enabled a GitHub setting which requires all commits to be signed, but we did not have all unsigned commits from OSBotify merged into production before enabling that setting. So we disabled the setting, merged that PR, and should be able to re-enable the setting without issues now.

[roryabraham]

I submitted a pull request to make the action fail loudly. Should we point our repo at the fork in the meantime?

[roryabraham]

No movement on the PR. We could still test with our fork. This is no longer critical because we figured out the root cause. Going to switch this to monthly.

[roryabraham]

Saw a tiny bit of movement on the pull request 3 days ago, I'm very hopeful that it will be merged soon.

[roryabraham]

This keeps biting us, I'm going to put up a PR to test my fork.

[roryabraham]

Got a PR up with an alternate solution – just ditch the community action and use the GH CLI!