What shipped, dated and specific. Big features land here; fixes ship quietly every week.
Run simulated incidents against your real runbooks, on a schedule or on demand. NOVA replays a recorded failure, executes the runbook in a sandboxed context, and reports where the steps drifted from reality.
Teams on the beta caught 14 stale runbooks in the first week. Drills never touch production state.
Grafana OnCall joins PagerDuty and Opsgenie as a first-class paging source. Grouping v2 correlates across sources, so a Datadog monitor and a Grafana alert about the same deploy now land in one incident instead of two.
You can now require approval for production while letting staging runbooks execute unattended. Policies live in your repo as YAML next to the runbooks they govern, and changes go through review like any other code.
The generated review doc now exports with the timeline, contributing factors, and action items intact, instead of a flat text dump. Action items sync back: close the Jira ticket and the postmortem updates itself.
Fix something live while NOVA watches the session, and it drafts the runbook: commands, guards, and verification steps included. You review and commit it like any other change. Most teams' first ten runbooks now start this way.