Increase E2E test granularity

[wrijeff]

Currently we have two end-to-end tests: one to test raw span processing and one to test the service map.

While adding new usecases (first peer forwarding, now trace group ES callbacks), we've updated the existing tests to cover new functionality. From #480, we should increase test granularity by splitting the current test into more feature-specific tests.

End state would be tests for:

1. Basic span test - Single DP
2. Basic service map test - Single DP
3. Peer Forwarder plugin test - Two DP with forwarding enabled
4. TraceGroup plugin test - Two DP without forwarding

etc.

This will make it easier to isolate breaking changes based on the specific test failing.