How fast should we complete a postmortem?

Draft within 2 business days, review by day 5, and close Sev1/Sev2 corrective actions within 30 days. Use waivers for exceptions with VP signoff.

How do we stay blameless but still hold people accountable?

Hold the system accountable for failure and leaders accountable for resourcing fixes. If an individual repeatedly bypasses guardrails, that’s a performance issue handled outside the review—not in it.

What tools do we need to start?

PagerDuty (or Opsgenie), Slack/Teams, a timeline bot (`incident.io` is great), Jira for tracking actions, Git for templates and runbooks, and Grafana/Prometheus for SLOs. Everything else is optional.

We’re ITIL-heavy. Will this fly with CAB and auditors?

Yes. Treat the postmortem as the Problem Record, map actions to Changes with approvals, and attach evidence (diffs, dashboards). Auditors prefer deterministic processes with artifacts.

Remote-first teams: any special advice?

Ritualize channels and roles, record the Zoom, and harden timelines. Make the paved path easy with ChatOps: one command creates channel, assigns roles, pins Zoom, and starts the timeline.

Culture · Oct 16, 2025 · 9 minute read

The Postmortem Ritual That Quieted Our 3 A.M. PagerDuty

Blameless doesn’t mean consequence-free. It means system-first accountability, leadership air cover, and a repeatable ritual that prevents the next page.

Riley Grant

Principal, GitPlumbers

Riley has led SRE and platform teams at fintech and retail orgs for 20 years—through monoliths, microservices, and AI-fueled on-call chaos. He now helps enterprises make production quieter without killing delivery speed.

“Blameless isn’t about letting people off the hook; it’s about putting the system on the hook.”

Back to all posts

The Postmortem Ritual That Quieted Our 3 A.M. PagerDuty

Key takeaways

Implementation checklist