PagerDuty
Connect PagerDuty to measure change failure rate and time to restore from real production incidents. Without an incident source, Antenna estimates both from git history, tracing each revert, hotfix, or rollback back to the deployment it repaired.
Prerequisites
- An enterprise admin role in Antenna
- A PagerDuty user with read access to the account you want to connect
- A connected Git provider, so there are deployments for incidents to be attributed to
Each PagerDuty account can be connected to one Antenna enterprise.
Setup
Go to Sources in Antenna
On app.antenna.dev, go to Settings > Sources. Click Add source and select PagerDuty from the dropdown.
Authorize Antenna
Click Connect, then Continue to PagerDuty. Sign in if prompted and approve read-only access to your account.
Wait for the import
You will be returned to Antenna where you can verify the connection is active. Incident data will begin importing in the background.
Data Collected
Using a read-only token, Antenna reads two main types of data from the PagerDuty REST API:
| Data | Details |
|---|---|
| Services | Your service directory, including the teams that own each service |
| Incidents | Incident number, title, status, urgency, priority, created and resolved timestamps, time to first acknowledgement, service, team, escalation policy, and who resolved the incident |
Antenna never collects alert payloads, incident notes, or log entries. By default, it collects high-urgency incidents across every service.
The initial import covers up to 18 months of incidents. After that, Antenna refreshes every few hours and re-checks the most recent seven days for new activity.
Metrics
With PagerDuty connected, Antenna measures two DORA keys from incidents instead of estimating them from git history:
- Change failure rate — deployments implicated by an incident, divided by successful deployments. An incident counts only once a person resolves it. Antenna attributes each incident to the last successful deployment that finished within 48 hours before the incident opened.
- Time to restore — the time from when an incident opens to when it resolves, averaged across implicated deployments.
The DORA report notes which incidents it uses to calculate each metric. Its drilldown shows the incident behind each failure, with a link back to PagerDuty.
Troubleshooting
PagerDuty shows a Reconnect button. The authorization was revoked or expired in PagerDuty. Click Reconnect on the source and authorize Antenna again.
Disconnecting. Antenna deletes the stored authorization, and change failure rate and time to restore revert to git-based estimates.