Closed bear-in-the-air closed 1 year ago
Self hosted runners cannot run jobs in parallel. The best we can do is register a second runner, which means that if two merge requests go up at the same time they do not need to queue side by side.
Because jobs are stateful, they cannot be deployed across multiple runners. Will raise a new low priority issue to investigate making CI jobs stateless.
As a daft engineer I don't want to wait long for my CI tasks to finish
Success criteria: