I ran into some more problems with Flyte. I registered a workflow locally with a remote Flyte cluster, then terminated it after a second, and the metrics kept hanging after that. When I set getmertrics depth=0 everything is fine, otherwise Flyte’s admin is killed because it accesses a null pointer and the whole cluster crashes, how should I avoid this?

@er Ksy can you elaborate a bit? Are you hitting this API?