Test stability lately worse than normal?

[joshtrichards]

Is it just me or are the tests particularly bad lately?

It's across the board but most notably:

• cypress <-- at first thought it was cache save/restore related but I see that part fails often and things still work so 🤷‍♂️
  • one key here to workaround: rerun all cypress jobs not just the failed ones (I think; or maybe this is only important for >24 hour old jobs)
• phpunit-mysql <-- gh workflow doesn't see the mysql container go healthy but it does go healthy from what I can tell
  • timing conflict between gh's health checks and the container ones?
• drone <-- group issues (maybe?)
  • random areas
  • silly stuff like being unable to resolve "github.com"
  • this one is harsh because many of us can't re-run drone so a rebase is required with a full set of tests all over again

[SystemKeeper]

• phpunit-mysql <-- gh workflow doesn't see the mysql container go healthy but it does go healthy from what I can tell

Seems the workflow was not updated to the template:
See https://github.com/nextcloud/.github/blob/master/workflow-templates/phpunit-mysql.yml and nextcloud/spreed#11627