Hi! I’m running into super weird behaviour that I don’t know how to debug any further - the logs don’t tell me what’s going on and I feel like I tried all the “shots in the dark”: I have a bunch of Airbyte syncs that I’d like to trigger. Doing them one-by-one is a bit slow, so I created a connection for each database, which allows me to trigger them in parallel in Dagster. However, doing so fails, with the following message:

Detected run worker status FAILED: 'K8s job failed'

When I switch to the in_process executor, it all works fine, but with the multiprocess executor (no matter the parallelism) it fails almost instantly (half a minute for the pod to spin up, then 5 seconds for the job to fail, without any info as to what or why)

cc @daniel

Ha, it seems to be an OOM issue on k8s that I thought I ruled out because of very generous resource limits. I just never apply the tags…