Fallthere status.
A check from outside looks at every service once a minute. Here you see what is working now and how the last ninety days went.
Services
Incidents
How we measure it
A program runs every minute on a computer that is not our server, so if the server stops, whoever is watching keeps watching. It tries the website, sign-in, the public pages and the connected services, and checks that the certificates are valid.
A service switches to “not responding” after two checks in a row go wrong, and back to “working” after two go well: a one-minute glitch changes nothing. “Slower than usual” means it responds, but takes more than three seconds.
Every day has its own color, on Rome time: red after five minutes of failed checks, yellow if there were a few minutes of trouble or a lot of slowness, grey if that day’s checks are not enough to say. The percentage is rounded down.
This page lives on the same server as the website: if the server is down it does not open, and the alert to whoever runs the system goes out from outside, without waiting for someone to notice.