TYPO3 Core

Our Bamboo plans are not in a good shape. We have code redundancy, lots of effort to overcome the mssql situation and random failures due to 3rd party issues.
To make the situation a bit better, we should revert the grouping of nightly jobs into shuffled batches to have at least big mssql blocks again. This way we can better detect whether we have real failures.

Additionally, cherry-picking can be retried up to n times which should also decrease the failure rate.