Skip to content

observability: add Grafana alert-overview dashboard (alerting plan ph… - #36

Merged
phonkd merged 1 commit into
mainfrom
worktree-alerting-phase1
Jul 19, 2026
Merged

observability: add Grafana alert-overview dashboard (alerting plan ph…#36
phonkd merged 1 commit into
mainfrom
worktree-alerting-phase1

Conversation

@phonkd

@phonkd phonkd commented Jul 19, 2026

Copy link
Copy Markdown
Owner

…ase 1)

Phase 1 of plans/alert-overview.md: a single provisioned dashboard giving one grouped "what's firing right now" page over the homelab alert rules, the OpsGenie-style overview the plan wants.

Built on the ALERTS series Mimir's ruler writes back (queried via the Mimir Prometheus datasource), not the native alertlist panel: alertlist is a known Grafana limitation and only renders Grafana's built-in Alertmanager, not the external Mimir Alertmanager datasource. ALERTS carries the same firing/pending state that feeds the Discord Alertmanager, so it sees every live alert.

  • stat row: firing critical/warning/total + pending
  • firing table + pending table, grouped/sorted by severity then host, severity colour-mapped (critical=red, warning=orange)
  • text panel noting it's read-only; silence/ack stays in Grafana Alerting

Zero blast radius: read-only provisioned dashboard on an existing datasource, obs host only. Deploy with deploy observability.

…ase 1)

Phase 1 of plans/alert-overview.md: a single provisioned dashboard giving
one grouped "what's firing right now" page over the homelab alert rules,
the OpsGenie-style overview the plan wants.

Built on the ALERTS series Mimir's ruler writes back (queried via the Mimir
Prometheus datasource), not the native alertlist panel: alertlist is a known
Grafana limitation and only renders Grafana's built-in Alertmanager, not the
external Mimir Alertmanager datasource. ALERTS carries the same firing/pending
state that feeds the Discord Alertmanager, so it sees every live alert.

- stat row: firing critical/warning/total + pending
- firing table + pending table, grouped/sorted by severity then host,
  severity colour-mapped (critical=red, warning=orange)
- text panel noting it's read-only; silence/ack stays in Grafana Alerting

Zero blast radius: read-only provisioned dashboard on an existing datasource,
obs host only. Deploy with `deploy observability`.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Copilot AI review requested due to automatic review settings July 19, 2026 14:06
@phonkd
phonkd merged commit a2c8ef0 into main Jul 19, 2026
2 checks passed

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Adds an OpsGenie-style “what’s firing right now” overview for homelab alerts by provisioning a Grafana dashboard that queries Mimir’s ALERTS series, plus a plan documenting the phased approach (Phase 1 dashboard now; optional Phase 2 Karma later).

Changes:

  • Add a provisioned Grafana dashboard (alerts.json) showing firing/pending alerts and summary stats from the ALERTS metric.
  • Add an implementation plan (plans/alert-overview.md) describing goals, trade-offs, and future optional Karma rollout.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 4 comments.

File Description
plans/alert-overview.md Documents the alert-overview approach/phases; currently contains a few remaining alertlist references that conflict with the implemented ALERTS-based Phase 1.
modules/grafana-dashboards/alerts.json New Grafana dashboard JSON providing stats + firing/pending tables sourced from ALERTS via the Prometheus datasource variable.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comment thread plans/alert-overview.md
Comment on lines +44 to +48
**Phase 1 (recommended first slice) — a Grafana `alertlist` dashboard.** Grafana is
already provisioned with dashboards-as-JSON (`modules/grafana-dashboards/*.json`,
declarative, read-only in the UI) **and** already has the "Mimir Alertmanager"
datasource. So the overview is *one new JSON file*: a dashboard whose main panel is
the native `alertlist` type reading firing alerts from that datasource, grouped by
Comment thread plans/alert-overview.md
Comment on lines +49 to +52
`severity`/`instance`. Zero new service, no traefik, no firewall — reachable at
Grafana `:3000` over wg-obs like every other dashboard. Trade-off: it's a *view*,
not a console — silencing/ack still happens in Grafana → Alerting, and it's VPN-only.
This covers the core "one page of everything firing, grouped" want.
Comment thread plans/alert-overview.md
Comment on lines +170 to +174
- **Grafana dashboard vs. Karma** — *recommend* **ship the Grafana `alertlist`
dashboard first** (Phase 1): it reuses infra we already run, is one declarative
JSON file, and covers the core "one grouped page of what's firing" want at
near-zero cost. Add **Karma** (Phase 2) only if that panel proves too thin — i.e.
you want inline silence/ack or a non-VPN internal URL. Not mutually exclusive; the
Comment on lines +29 to +32
"options": {
"mode": "markdown",
"content": "**Everything firing right now**, read from the `ALERTS` series Mimir's ruler writes back (same rules that feed the Discord Alertmanager). This is a read-only view — to **silence or ack**, go to Grafana → **Alerting → Silences** on the *Mimir Alertmanager* datasource. Grafana's native `alertlist` panel can't render an external Alertmanager, so this uses the `ALERTS` metric instead."
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants