Overhaul stats: start collecting stats per server instance (per socket)

[josecelano]

Start collecting stats per server instance (per socket).

Sub-tasks

• HTTP Tracker Instances.
  • Instantiate independent services (event sender and stat repo) for each instance.
  • Add E2E tests for global metrics (it should fail after the previous refactor).
  • Fix the failing tests (add a listener for the instance to update global metrics).
• UDP Tracker Instances.

[josecelano]

This commit creates services per HTTP tracker instances (server bound to a socket address) to collect metrics only for that instance.

There should be test failing because each HTTP tracker instance can use only one event sender and metrics repository. Therefore global metrics should not be updated after this commit.

I think the problem is there are no E2E tests for statistics. There are only integration tests at the HTTP Core package level. That means metrics per instance work.

I will add an E2E tests for global metrics even if we are planning to remove them in the future. In the mean time, it will be covered.

cc @da2ce7

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 45 lines in your changes missing coverage. Please review.

Project coverage is 81.83%. Comparing base (734393d) to head (a9b1c14).

Files with missing lines	Patch %	Lines
src/container.rs	0.00%	43 Missing ⚠️
src/console/profiling.rs	0.00%	1 Missing ⚠️
src/main.rs	0.00%	1 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #1404      +/-   ##
===========================================
- Coverage    82.00%   81.83%   -0.18%     
===========================================
  Files          232      232              
  Lines        17046    17082      +36     
  Branches     17046    17082      +36     
===========================================
  Hits         13979    13979              
- Misses        2831     2867      +36     
  Partials       236      236              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[josecelano]

There is another problem. You can not halt the application with CTRL+C. I think it might be a deadlock. Only the servers before the HTTP servers are started. I'm going to try a different approach. In the AppContainer initialization I initialize all the services needed: the global ones and the duplicates ones that we need for each tracker instance. This way the start function would be pretty much the same.