Hi @Malthe,

Good callout on Spark UI access. You are right. The "Open in Spark UI" link is only documented and supported for Databricks SQL queries on SQL warehouses, not for serverless notebooks or jobs. So for pure serverless jobs/notebooks, you don’t currently have that fallback to the classic Spark UI, which makes the gaps in the query profile UI even more visible.

I agree this still falls short of the "every billed minute is traceable" experience you are used to from classic Spark UI, especially on serverless. Your feedback about making the serverless profiling view reconcile cleanly with DBUs (and exposing more of the hidden/grouped work) is completely valid and is being passed on to the product team.

If this answer resolves your question, could you mark it as “Accept as Solution”? That helps other users quickly find the correct fix.