Thanks for clarifying. This also highlights the lack of sufficient testing, and subpar processes, eg automation should be tested against critical paths, and changing a license should require human approval.
Agree. The lack of tests has been the biggest regret of mine for this project. It started as a hackathon project a long time ago, and as it grew up, it never got the testing investment it deserved. I imagine the code coverage is about 0.001% and there's no end-to-end tests in place.
My local tooling doesn't fail as visibly as this but it certain fails just as catastrophically leaving me feeling excessively silly.