Incident Management

Incidents communicate downtime on one or more components. You can open incidents manually as part of your incident management process, or automatically with automation rules. When an incident is posted, a notification will be sent to all users subscribed to the Status Page, unless you opt out of notifications when creating the incident update.

​

Creating an incident

Create an incident when you want to communicate downtime or known issues to users via your Status Page.

1. Click ‘Status pages’ in the sidebar and open your Status Page.
2. Click ‘Add incident’ at the top right.
3. Enter the name of the incident.
4. Select which components are impacted, and the impact each one gets: under maintenance, degraded performance, partial outage, or major outage.
5. Write the initial incident update message, and set the status of the incident.
6. If you want to backfill the incident, select Use custom date. See Backfilling incidents for more information.
7. Select if you want to notify subscribers of your Status Page about the incident.
8. Click ‘Create incident’

The incident now appears on your status page, and the impacted components show their new status. If you selected the option to notify subscribers, they receive the update message by email.

​

Incident duration and downtime

Downtime is tracked per component, not per incident. Each affected component has an impact timeline made of phases:

• A phase starts when you give a component a non-operational impact, and ends when you change the impact, set it back to operational, remove the component from the incident, or resolve the incident.
• Changing the impact (e.g. major outage → partial outage) closes the current phase and starts a new one, so the full history is kept.
• A component can only be non-operational while an active incident impacts it. Resolving the incident closes all open phases.
• Phases of an ongoing incident count up to the current time.

Uptime is calculated from these phases over the last 90 days. Phase time is weighted by impact: Overlapping phases never double-count: when several incidents impact the same component at the same time, the overlapping time counts once, and partial outage time that overlaps a major outage is already covered by the major outage. For example, a component with a 2-hour major outage and a separate 10-hour partial outage in the 90-day window accumulates 2 + (0.3 × 10) = 5 hours of downtime. You can correct the record after the fact by editing a component’s impact timeline on the incident, or by backfilling an incident. Edited timelines must stay consistent: a phase ends after it starts, phases of the same component never overlap, and a resolved incident has no open phases.

​

Incident notifications

Whenever you post an incident update, Checkly can automatically send email notifications to users subscribed to your Status Page. Emails are sent out only when an update is first posted. Notifications are not sent again if the incident update is edited. If you want to post an incident update without notifying your subscribers you can uncheck the ‘Notifications’ checkbox when updating the incident. Incident notifications contain the incident name and status, and the impacted components with their impact.

​

Subscribing to incident notifications

Subscribing to incident notifications is done on your Status Page. To subscribe, a user needs to enter a valid email address. A verification email will be sent to the registered email, and the user will be ready to receive incident notifications after confirming the subscription.

​

Incident automation

Automation rules open incidents automatically when checks fail. A rule matches checks by tag: when a check carrying one of the rule’s tags fails and triggers an alert, the rule opens one incident impacting the components it lists. The incident resolves automatically when the check recovers. To create an automation rule:

1. Open your Status Page and go to the ‘Automation’ tab.
2. Click ‘Add rule’ and name the rule.
3. Add one or more tags. Any check or group carrying one of these tags can trigger the rule.
4. Select the components the incident should impact, and the impact each one gets.
5. Write the first status update (posted when the incident opens) and the last update (posted when it resolves).
6. Select if subscribers should be notified, and optionally adjust the cool down between automated incidents.
7. Save the rule.

Automation rules can also be managed in code with the StatusPageV3AutomationRule construct.