Summarize Trace data by service and operation

[copa2]

Requirement - what kind of business use case are you trying to solve?

As user I want to see where do I loose time in a big trace. This would allow me to better understand where I can optimize.

Problem - what in Jaeger blocks you from solving the requirement?

Tracepage already provides a lot of insights. But it gets a bit overwhelming when viewing traces with 1000 spans.

Proposal - what do you suggest to solve the problem or improve the existing situation?

Add a component which summarize data of a trace:
Count, TotalTime, % of time grouped by service and operation.

NOTE: There is already #158. But it is quite large in context.
This is probably a part of it. I decided to use totalTime instead of Avg. time. It helps more to see which are the big chunks. But we could also add avg. time additionaly to it.

[copa2]

Created a POC for better explaining. If this is of interest I will create a PullRequest.

• Count: 286 calls on this service.operation
• TotalTime: 9951 ms in this service.operation (sum of 286 span.durations)
• % of trace: 18.25 % time of the whole trace duration is spent in this service.operation
• InMethod: 737 ms time spent in this service.operation (Takes TotalTime and subtracts all CHILD_OF durations).

Example of a trace:

[yurishkuro]

@copa2 this looks awesome! A PR would be great. A couple of notes on the UI:

• once you switch to trace graph, I think it would be useful to keep the trace header the same as on the Trace View (except for minimap). It would also allow switching back to the Gantt chart view or other views. In fact, in the long term it would be great to have those views toggled without reloading the page.
• have popups explaining what each number means