Open an incident
Go to Incidents and use the creation control. Select the affected endpoint, enter a clear title and start time, choose a severity, and add an initial update describing observed impact. Use the timezone shown by the app when entering times.
| Severity | Suggested use |
|---|---|
| Critical | A major service capability is unavailable and needs immediate response. |
| Major | Significant customer impact or an important service disruption. |
| Minor | Limited impact or partial degradation. |
| Informational | An event worth recording without the same outage urgency. |
These are communication choices for your team; severity alone does not replace alert-route configuration.
Keep the timeline current
Open the incident detail page to review its timeline and monitoring evidence. Add updates that separate confirmed facts from investigation. Use Investigating while determining the cause, Ongoing while the incident continues, and Resolved once recovery is confirmed.
A useful update states the affected service, observed impact, current action, and what you will check next. Avoid copying credentials, private customer data, or sensitive internal URLs into a message that may be shared publicly.
Resolve and review
Confirm recovery with recent checks, update the incident status, and record the resolution reason. Acknowledging an on-call alert is separate from resolving an incident: acknowledgement assigns response responsibility; resolution records the service outcome.
Use the post-mortem controls to record the cause, timeline, and follow-up actions, then publish only the material intended for your audience. Pro includes PDF incident reports for sharing or archiving.
Public communication
Public incident views support communication with visitors. Review the published view and its content before distributing a link. A public status page also depends on the organization's publishing setting and the affected monitor's public visibility. See status pages and sharing.
Automation
The Management API lists incidents and updates existing incidents. The current collection route does not support creating incidents with POST; use the application to create them. The API reference lists the supported methods.