Manage Incidents and Write Postmortems

Get through production incidents with less chaos, and make sure the same one does not come back.

What it does

• Turns alerts into a first diagnosis for the on-call person: what is failing, impact, likely severity, evidence, first checks
• Groups related alerts, drafts the incident kickoff, pages and status updates
• Looks for what changed (deploys, config, traffic) to point at the likely cause
• Posts a daily on-call briefing: who is on call, open incidents, overdue follow-ups
• Writes blameless postmortems with a sourced timeline, root cause and owned action items
• Finds recurring causes and open actions across past incidents

What you need

• Nothing to start: paste an incident thread or try it on a sample incident
• PagerDuty, to read alerts, incidents and the on-call schedule and page responders
• Slack, to read incident channels and post updates (Opsgenie, Teams or another tool can be connected instead)

How setup works

A postmortem arrives in the first minutes. Along the way it learns your severity levels, services and owners, connects PagerDuty and Slack, triages a real alert and confirms the daily briefing.

What stays in your hands

• It drafts pages, announcements and tickets and sends them only with your go-ahead; the daily briefing is the one message it posts on its own, once you approve it