Improve metrics for Aggregate streams #18323
Description
Is your feature request related to a problem or challenge?
Currently, aggregate streams have some of the least descriptive metrics in the framework.
(only baseline metrics are available).
This makes it hard to identify bottlenecks and query planning issues, especially when running on remote or distributed systems.
Describe the solution you'd like
Some more detailed metrics along with the baseline metrics.
e.g., how long did it take to calculate the groups? how long did the aggregate arguments take to evaluate? how long did the grouping accumulators work for? and how long did constructing the batches take.
Describe alternatives you've considered
I've added some instrumenting to local runs, but this is far from providing good statistics when releasing.
Additional context
No response