We have a lot of great merge requests in the queue. Even though it would be great to get them all in right now, I propose to take them one at the time and let the pipeline go green before merging the next so we don’t have to engage in forensics trying to figure out what caused what errors and how to fix them. If something goes wrong, I suggest using the
stop-the-line principle. Basically:
- No more merges or pushes to master other than trying to fix the problem.
- If you can’t fix the problem rather quick:
- Back out the changes
- Get a green build/test
- Allow merge/push again
What do you think?