The number of random failures is growing.
In an ideal world we'd fix these failures, but as a stop gap to avoid rerunning multiple jobs multiple times it would be great to group these jobs so most tests can pass or fail on their own changes.
Open an MR with a trivial change.
See several functional failures
Mark them as @random
This will group them into their own section.
Ideally they would be split one to a job in that section with some sort of retry. I'm not sure if this is feasible.
Active
11.0 🔥
phpunit
For those tests with real random-generated values, it would help if these values were reported, so that hopefully it can be reproduced with exactly these values.
One idea could be to pass random values from a data provider, but perhaps there are easier solutions.