Is there a way to have Dagster automatically kill a stuck step (specifically from a DynamicOutput)? We have tasks that query an external database, and for some reason they're intermittently hanging; we're working separately on that. But at the orchestration level, it would be cool if we could set a timeout on per-op execution for these steps and have Dagster automatically kill them (and maybe retry which I know Dagster can do).
07/19/2022, 5:12 PM
Hi @Mark Fickett this isn't something we support right now, but i'll open a GH issue