Component(s)

No response

Is your feature request related to a problem? Please describe.

Traces hold information about all events happening in a system. Thus if we could convert traces into profiles, we could see a flamegraph of all processes happening over all components in a centralized view and would help to debug bottlenecks on macro level.

Describe the solution you'd like

A connector converting spans into profiles. Such connector would work great together with collector balancing by traceId + tail sampling.

Describe alternatives you've considered

Alternatives to this would be to make a specific trace storage/vendor to support this feature, which sounds a bit excessive since we could potentially utilize profiling systems for this.

Additional context

I would be happy to try building it, if you would consider it useful/viable.

Cheers!