proton-bridge: add Prometheus metrics endpoint

[mgarratt]

Summary

• add a new supervised metrics service in the proton-bridge image
• serve Prometheus metrics at /cgi-bin/metrics on CONTAINER_METRICS_PORT (default 9154)
• export service/process/probe/restart metrics based on existing s6 and healthcheck signals
• wire new service/files into the Dockerfile and require CONTAINER_METRICS_PORT at startup
• extend healthcheck to verify metrics service and endpoint readiness
• document metrics and current account sync-status limitation in README

Validation

• sh syntax checks for changed scripts
• ./scripts/lint.sh (shellcheck + hadolint)

[mgarratt]

Superseded by #4, which scrapes Bridge's gRPC API for real per-account login/auth state instead of external socat/CGI probes. Closing in favour of that approach.