feat(observability): add Grafana alerting rules, dashboard, and SLO provisioning (story 7-11)
Story 7.11: Observability SLOs
Meta-monitoring for the observability stack — "who watches the watchers".
Changes
- 4 alert rules: OTel Collector Down (critical), VM Storage High, VLogs Ingestion Drop, VTraces Export Failures (warning)
- Dashboard: Observability Stack Health with 8 panels (ingestion, storage, latency, health checks)
- Notification routing: contact point + notification policy
- Infrastructure: uid added to vm-datasource.yml, Ansible vars
- Tests: 63 unit tests — all pass
- Docs: retention policy + alerting documentation
Closes #820 (closed)