CFR from PagerDuty

Incidents as failure signal. Correlate to deploys via a 60-minute window.

Oliver Wakefield-Smith

Founder, Digital Signet. Built siemcost / monitoringcost / itbudgetcalculator.

Updated 21 Jun 2026

Direct answer

PagerDuty Incidents API exposes created_at, resolved_at, service, and urgency per incidentPagerDuty Incidents API2026. Correlate to deploys within a 60-minute window and tag matched deploys as failures. Exclude externally caused incidents (third-party, upstream) via custom-field convention PagerDuty DevOps Metrics2024.

The correlation window

60 minutes is the operational default. Smaller windows (15-30 min) miss slow-burn failures. Larger windows (4+ hours) attribute failures to deploys that did not cause them. Tune to your deploy cadence: high-DF teams can use 30 min; low-DF teams may need 4 hours.

External-cause exclusion

PagerDuty allows custom fields per incident. Use a root_cause: externalfield on incident-resolution to exclude third-party-caused incidents from CFRAtlassian DORA2024.