Alex

We're hoping to take this fix on this sprint.

My team and I are making good progress. We have a functional POC. It just needs some polish. It involves sending the run execution date to the logging handler since...

@thesuperzapper, if I understand the logic correctly, the handler does in fact try to get the log artifact path from the workflow status **first** (if the k8s pods have been...

@kromanow94 Thanks for the context! > I take that this is the assumption because it wouldn't be too good to have a requirement of storing the Argo WF CR Objects...

> if the log artifact URI is not saved in the MLMD, do we have any other place where we can store that information, or currently we're forced to guess...

Another argument in favor of storing the log as an output artifact is what if an admin elects to update the keyFormat? The logs for historical runs will no longer...

I'm starting to lean in that direction as well, but what the heck do I know, I am just a ~plebe~ user. Do ya'll think it make sense to submit...

IMO, the fundamental gap is the lack of an SDK. Data scientists would rather write python than yaml (for good reason). There needs to be (a) some clarification (and documentation)...

@charlesa101 that's json with no customization, and the configuration options are [abundant](https://spark.apache.org/docs/latest/running-on-kubernetes.html#configuration). It's nice to be able to just use ResourceOp though. Thanks for sharing. Our platform provides both pyspark...

Our organization prohibits refresh tokens with OIDC to optimize security. We would greatly appreciate it if kr8s didn't assume the presence of a refresh-token field in the Kubernetes configuration file,...