Status page
/status is the page your customers open when something feels broken. It answers in a sentence: "Payments are failing. The shop and the API work." Below it, one grid shows the last 90 days with a row per component, and a logbook keeps every incident in the words you wrote.
Choose what it shows
On the Status page screen: pick monitors and alerts, name them as customers would, group them, and see the live preview. Or in code, with defineStatusPage, where it is reviewed and deployed like the rest. When the file exists, the screen shows it read-only.
Only what is listed shows. A monitor that goes down marks its component on its own. An alert does the same for anything it watches: an error rate, a stuck queue, failed payments.
Incidents
Declare an incident on the Status page screen with a title, the affected components and its impact (minor, major, critical or maintenance), then post updates as you learn more: investigating, identified, monitoring, resolved. Each update shows on the page at once, in the logbook.
Agents can do it too, with a write token:
declare_incident({ title: "Payments failing", impact: "major", components: ["alert:failed-payments"], message: "Card payments are failing. Orders already paid are not affected." })
post_status_update({ incident: 12, state: "resolved", message: "Card payments work again." })
Visitors read titles and messages as written, so keep them in plain words. The page is private until you make it public, on the screen or with public: true in code.