When running multiple pipelines at same time (like in the order of 50-100, each containing 10 jobs), the database encounters concurrency problems and leaves the jobs in running state, though they actually failed because problems with the database were found.
I consider the database should be able to cope with that level of concurrency, but otherwise, the state of the database should at least be keep coherent.
When running multiple pipelines at same time (like in the order of 50-100, each containing 10 jobs), the database encounters concurrency problems and leaves the jobs in running state, though they actually failed because problems with the database were found.
I consider the database should be able to cope with that level of concurrency, but otherwise, the state of the database should at least be keep coherent.