archetech / archon / 24317022038

95%

main: 89%

Pull #404

github

Emit process_* metrics from Rust gatekeeper for Grafana parity

The existing archon-services Grafana dashboard queries
process_resident_memory_bytes and process_start_time_seconds to render
the gatekeeper Uptime and Memory panels. The TypeScript service got
these for free from prom-client's collectDefaultMetrics; the Rust port
emitted neither, so those panels went blank after cutover.

Enable the `process` feature on the prometheus crate and register
ProcessCollector::for_self() on Linux, which exposes:
  process_cpu_seconds_total
  process_open_fds
  process_resident_memory_bytes
  process_start_time_seconds
  process_virtual_memory_bytes
Gated with #[cfg(target_os = "linux")] because the collector reads
/proc — all our containers are Linux so this is a no-op elsewhere but
keeps cross-compiling clean.

nodejs_heap_size_used_bytes stays absent — Rust has no JS heap and
process_resident_memory_bytes is the correct substitute.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

1990 of 2192 branches covered (90.78%)

Branch coverage included in aggregate %.

4231 of 4345 relevant lines covered (97.38%)

1167.44 hits per line