Coming soon — stub
This page is a placeholder. The monitoring VM and dashboards haven't
shipped yet — see .claude/features/deploy-pipeline/plan.md §
P4 for the implementation plan and status.
What this page will cover
- Topology: dedicated monitoring VM with Prom + Grafana +
Alertmanager;
node_exporter+cAdvisorquadlets on every app host. - Scrape config: file-based service discovery, one target file per
host, generated from
inventory.yml. - Adding a new metric source: PR shape against
shyft-infra/monitoring/. - Dashboards-as-code: where the JSON lives, how Grafana file-provisioning picks them up, the starter dashboards (system, podman containers, per-app).
- Alert routing: Alertmanager → Slack for v1, future routes (PagerDuty, email), template conventions.
- cAdvisor + rootless podman gotchas — the cgroup mount and permissions the container needs.
Until then
The shape is sketched at § Telemetry flow and § Monitoring VM on the parent page.