#What's Changed

[Feature] A new incident history page lists every outage and degraded period from the past 90 days, grouped by month, with the affected service and when it started and ended. Filter it to the services you care about, step back three months at a time, and open it from "View incident history" under Past incidents on the status page.
[Improvement] The database response-time chart on the status page is interactive: hover or tap it, or use the arrow keys, to read the average for each half hour.
[Fix] An incident that is still happening is now labelled ongoing reliably. Checks run every 5 minutes, and the page used to treat a problem as over when the last failed check was more than 2 minutes old.