Incidents

Lifecycle

opened ──▶ investigating ──▶ identified ──▶ monitoring ──▶ resolved
                                                │
                                                └──▶ reopen → monitoring

• Opened — created. Auto or manual.
• Investigating — default initial status. Customer-visible if you publish the incident to a status page.
• Identified — root cause known, fix in flight.
• Monitoring — fix deployed, watching for regressions.
• Resolved — closed. Resolution summary is optional but encouraged; it surfaces in the postmortem.

Auto vs manual

• Auto — the alert engine opens an incident when a rule’s condition stays matched across the configured threshold. Severity comes from the rule. Auto-resolves after N consecutive OK checks (3 by default).
• Manual — declare from the topbar + → Declare incident. Pick severity, optionally attach affected monitors, write a title.

Severities

• P1 — full outage, customer-impacting, all hands.
• P2 — significant degradation; alert your on-call.
• P3 — partial / regional issue, no immediate customer impact.
• P4 — minor; investigate during business hours.
• P5 — informational; no action required.

The numbers don’t enforce policy; they’re a shared vocabulary.

Acknowledging

Press a on an incident detail page to acknowledge. Acks record who’s looking and silence subsequent alerts on the same incident. Doesn’t change status — you still drive that explicitly.

Resolving

Click Resolve, or press Shift+R. The resolve dialog asks for an optional summary that’s stored on the incident and rendered into the postmortem.

Comments

Every action on an incident — status change, severity change, ack, public update, resolve — appears in the activity timeline. You can also add free-text comments (press c to focus the composer).

Postmortems

Visible only after resolved. Free-form Markdown. Renders inline using Pulse’s small Markdown renderer (headings, code, lists, links — no script, no embedded HTML).

Tip

Pulse does not enforce a postmortem template. Use whatever your team’s process expects. We recommend at minimum: impact, timeline, root cause, remediation, follow-ups.

Linking to status pages

From an incident’s detail page, Send public update posts to a status page tied to the incident. The update lands in the active-incident timeline on the public page. See Status pages → Linked incidents.