Not every issue is a sev-1. A clear severity rubric prevents overreaction and underreaction at the same time.
Tie severity to user impact and release scope, not to anxiety.
Define severity levels
- Sev-1: widespread outage or data loss.
- Sev-2: partial outage or major feature broken.
- Sev-3: degraded experience or limited cohort impact.
Give examples
Examples reduce debate. Write one example per severity level so the team can align quickly.
Escalation paths
Each severity level should map to an escalation path: who gets paged, who communicates, and how quickly.
Link severity to the release note
If a release caused the incident, link the severity assessment in the release log. This creates traceability.
How ReleaseMind helps
ReleaseMind links incidents and releases so severity context stays attached to the release narrative.
