starknet_transaction_prover: HTTP request count + latency + in-flight metrics

[avi-starkware]

Adds HttpMetricsLayer placed between the monitoring endpoints and the
rest of the stack so probes/scrapes don't distort the request-latency
distribution. Records request count by bounded method/status labels,
end-to-end latency, and in-flight gauge. Also adds a shared test recorder
helper that installs the global Prometheus recorder once across the
test binary.

Co-Authored-By: Claude Opus 4.7 (1M context) noreply@anthropic.com

[cursor]

PR Summary

Low Risk
Observability-only middleware with bounded labels; layer ordering is documented and covered by tests, with no changes to proving or auth paths.

Overview
Adds HttpMetricsLayer to the shared prover_http_middleware! stack so JSON-RPC traffic gets Prometheus request counts, end-to-end latency histograms, and an in-flight gauge, with bounded method / status labels and no user-controlled cardinality.

The layer sits below HealthLayer and MetricsLayer, so GET /health and GET /metrics are not counted in those series. Startup calls preregister_http_metrics and configures HTTP_DURATION_BUCKETS for prover_http_request_duration_seconds; in-flight tracking uses a new GaugeGuard RAII helper in metrics.rs. README documents the three new metrics; unit tests assert recording and probe/scrape exclusion.

^{Reviewed by Cursor Bugbot for commit 17f4f71. Bugbot is set up for automated code reviews on this repo. Configure here.}

[avi-starkware]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• starknet_transaction_prover: log resolved config at startup (secrets redacted) #14684
• starknet_transaction_prover: origin-level breadcrumbs for validation errors #14172
• starknet_transaction_prover: /health returns 503 when service is saturated #14171
• starknet_transaction_prover: track saturation and queue depth on the request path #15027
• starknet_transaction_prover: bump prover_panics_total on panic hook fire #14170
• starknet_transaction_prover: HTTP request count + latency + in-flight metrics #14169 👈 (View in Graphite)
• starknet_transaction_prover: proving-job duration + outcome metrics #14168
• starknet_transaction_prover: Prometheus /metrics endpoint with build_info #14167
• starknet_transaction_prover: extract the TLS accept loop and test its drain #15049
• starknet_transaction_prover: graceful shutdown on SIGTERM/SIGINT #15026
• starknet_transaction_prover: global panic hook with structured log #14166
• starknet_transaction_prover: document existing logging, request ids and health probe in README #14203
• main-v0.14.4

This stack of pull requests is managed by Graphite. Learn more about stacking.

[reviewable-StarkWare]

This change is 

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit ce52272. Configure here.}

[cursor]

Flaky absolute in-flight gauge assert

Low Severity

The in-flight check uses an absolute scrape value of 0.0, while this file’s own header notes that the Prometheus recorder is process-global and other tests (including ones that drive HttpMetricsLayer) run in parallel. A concurrent in-flight request can make this assertion fail intermittently, and a steady 0 also does not prove GaugeGuard ran because preregistration already sets the gauge to zero.

 

^{Reviewed by Cursor Bugbot for commit ce52272. Configure here.}