[metrics]: Add Prometheus metrics module

[pseudomuto]

Introduce internal/metrics, an fx module that constructs a hashicorp go-metrics instance backed by a Prometheus sink and serves it at /metrics over HTTP. The server is bound to the fx lifecycle: it starts in a background goroutine on OnStart and shuts down gracefully on OnStop, bringing the whole app down with a non-zero exit code if it stops for any reason other than a clean shutdown.

Wire the module into the serve command and add tests covering provider construction, an end-to-end /metrics scrape across the lifecycle, and the missing-address failure path.

[Copilot]

Pull request overview

Adds an internal/metrics fx module that wires HashiCorp go-metrics to a Prometheus sink and exposes it via an HTTP /metrics endpoint, then integrates it into the proxy serve command and validates behavior with lifecycle-focused tests.

Changes:

• Introduce internal/metrics fx module to construct a Prometheus-backed go-metrics instance and serve /metrics via an HTTP server bound to the fx lifecycle.
• Wire the metrics module into cmd/proxy/serve.go with a new --metrics-addr flag.
• Add tests covering lifecycle scrape behavior and failure/shutdown behavior.

Reviewed changes

Copilot reviewed 5 out of 6 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
internal/metrics/fx.go	Adds the metrics fx module, Prometheus sink construction, and lifecycle-managed HTTP server.
internal/metrics/fx_test.go	Adds end-to-end tests for /metrics scraping across Start/Stop and listener-failure shutdown behavior.
internal/metrics/doc.go	Adds package-level documentation for the metrics module.
cmd/proxy/serve.go	Wires the metrics module into the serve command and adds CLI/config plumbing for the listen address.
go.mod	Adds github.com/hashicorp/go-metrics dependency (and indirect deps).
go.sum	Records checksums for newly added module dependencies.