I have a job using the k8s Run Launcher, and Multiprocess Executor, which often fails because of a

DagsterExecutionInterruptedError

. This isn’t because of a termination request; it’s just a mysterious failure because something is killing my process. Maybe it’s out of memory? I’d like advice on how to debug this sort of thing - how can I get more inspection on the reason for a failure like this?

Hi Spencer, I created a gh discussion with some debug steps https://github.com/dagster-io/dagster/discussions/12943

Thanks. It turns out the pod was evicted because the node was low on ephemeral storage. This took a lot of investigation to figure out; it would be terrific if dagster-k8s could present more information on failure. I’ll see if I can come up with a more concrete feature request.

➕ for surfacing more debug info to the UI. Something that’s on our plate for sure