Add First Full-Round Metrics Test Case

[jutaro]

Summary

This PR introduces an end-to-end test for the metrics pipeline, covering everything from tracing a metric to scraping it via the Prometheus endpoint and validating the result. It is the first full-round verification of our metrics infrastructure.

Motivation

Until now, our metrics tests were mostly unit-level or partial integration checks. This new test ensures that:

• The Prometheus simple server actually serves the metrics we trace.
• Our tracer configuration and formatter properly register and emit metrics.
• The scraping path produces the expected value, closing the loop between producer and consumer.

This increases confidence in the observability layer, especially ahead of future instrumentation work.

What the Test Does

• Thread 1

  • Starts a Prometheus simple server bound to localhost:9090.
  • Sets up an ekgTracer and configures it.
  • Emits a Measure 42 metric.
  • Spawns Thread 2.
  • Waits for completion or shutdown signal.

• Thread 2

  • Scrapes http://localhost:9090/metrics.
  • Parses the metric output into a map.
  • Looks up the "measure" metric and asserts its value is 42.
  • Fails early if the metric is missing or has an unexpected value.

• On success, prints Got correct metric value ✔.