Improved visibility/workflow for run failures and re-executions

[chazmo03]

What's the use case?

As an operator of a Dagster deployment, I cannot find a decent way to ensure that all of my failed runs have been addressed. This is particularly challenging in the case where there are many failures, e.g., if a database was down, or a bug was deployed.

One can clearly see that there are a bunch of failures in the Runs > Failures tab. But, after having fixed the underlying issue that caused the failures, there is no way to reliably re-execute all of the failures (you can only re-execute the 30 per page as I see it). And after restarting the runs, there is no way to show that the original failed run was re-executed, since at that point the concern is now on the re-executed run.

Ideas of implementation

One way to improve this could be to have an additional status for re-executed runs, so when you re-execute a failed run, then status changes from Failure to something like Re-executed, and then the Re-executed run drops out of the Runs > Failures tab.

Additional information

No response

Message from the maintainers

Impacted by this issue? Give it a 👍! We factor engagement into prioritization.