Errors get lost in logs when one run fails on an experiment with multiple runs per pod

[taufeeque9]

The pod also gets a completed status which makes it difficult to check if any of the runs failed or not.