When you re-run a pipeline the logs and artifacts for those runs are seemingly lost forever. It would be great if we could look at previous runs outputs even after re-running a pipeline or step. This would make things like raising concerns with a specific pipeline run or flagging flaky tests a lot easier.