feat(metrics): add low-cardinality mode for prometheus metrics

[umputun]

Summary

• Add --mgmt.low-cardinality flag to use route patterns instead of raw paths for http_response_time_seconds histogram labels
• Reduces metrics cardinality when proxying services with dynamic URLs (e.g., /api/users/123 becomes ^/api/users/(.*))
• Falls back to [unmatched] for requests that don't match any route

Closes #160

[Copilot]

Pull request overview

This PR adds a low-cardinality mode for Prometheus metrics to reduce metric explosion when proxying services with dynamic URLs. The feature is opt-in via the --mgmt.low-cardinality flag, which switches the http_response_time_seconds histogram from using raw request paths (e.g., /api/users/123) to route patterns (e.g., ^/api/users/(.*)) as labels.

Key Changes

• Added MetricsConfig struct with LowCardinality option to control metrics label behavior
• Updated NewMetrics to accept configuration and use route patterns from request context when enabled
• Added --mgmt.low-cardinality CLI flag to enable the feature

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
app/mgmt/metrics.go	Core implementation: added MetricsConfig, updated NewMetrics signature, added getRoutePattern helper to extract route patterns from context
app/mgmt/server_test.go	Added comprehensive tests for low-cardinality mode and updated existing tests to use new NewMetrics signature
app/proxy/proxy_test.go	Updated all test cases to pass empty MetricsConfig{} to NewMetrics
app/proxy/health_test.go	Updated test cases to pass empty MetricsConfig{} to NewMetrics
app/main.go	Added LowCardinality flag to management options and wired it to metrics initialization
README.md	Added documentation explaining the low-cardinality feature and its benefits

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.