In some cases, it could be convenient to simply re-run the platform that caused an integration failure for a change to help investigate flakiness or test new fixes. Having this option could reduce CI resource usage, especially if suggested as a default to the user when triggering a precheck if the change has previously failed an integration.

Issues like latency for collecting COIN workitem data needs to be investigated, along with a complexity assessment need to be addressed.