Add relay Service Bus Slack alert monitor

[linear]

Goal

Add internal Slack alerting for the relay Service Bus topic so we are notified when relay delivery is unhealthy.

Scope

• Monitor bildrelaybus / relay-events subscriptions: bnj-dev, bnj-prod, op-dev, op-prod.
• Alert in #eva-ops-alerts when a subscription has dead-letter messages.
• Alert when the oldest active message is older than 30 minutes.
• Keep the implementation simple and stateless: every hourly run may report the same still-open issue again.
• Avoid Blob-backed alert state, dedupe storage, and recovery notices.
• Do not include message bodies, SAS URLs, connection strings, or Slack secrets in alert payloads.

Verification

• Unit coverage for stale-active alerts, dead-letter alerts, stateless repeated reporting, missing Slack config, and monitor failure handling.
• Service Bus smoke test with a temporary relay-alert-smoke subscription.
• Live settings verifier confirms exactly one monitor host is enabled.

Related

PR: https://github.com/bild-engineering/bild-ia/pull/299