Faster feedback for long traces

[doppleware]

Why are we building this?
Currently, Digma has one pipeline that requires a trace to end for the processing phase to start, this means users are delayed in getting feedback about their traces, especially if the traces are long.

What are we building?
Separate the feedback pipeline (showing up in the observability panel) vs. insights processing that will depend on trace being completed. This will result in faster feedback times for long traces