RTPengine Naming Drift Was a Monitoring Bug, Not a Media Bug

Monitoring configuration is production code. False alerts deserve root-cause treatment because noisy gates reduce operator trust and increase the chance that a future real warning is bypassed. The media relay was running under voip-platform-rtpengine, while the change-window script still searched for rtpengine. The resulting failure looked like a production dependency outage even though media service state was fine. Operational tooling had a hard-coded identifier that no longer matched the runtime source of truth. This is observability drift.

The sentinel name was corrected immediately and committed as its own fix so the safety gate reflected the current deployment. Where possible, derive runtime identifiers from reviewed service catalogs or Compose metadata instead of maintaining duplicate hard-coded lists. The concrete hserver evidence is commit dfe9992, so this note is tied to an actual production change rather than a hypothetical failure.