All systems operational
Every service we check answered normally the last time we asked.
What we check
Each bar is 30 minutes, over the last 24 hours
Database
OperationalThe store behind every board, message and meeting in your workspace.
Up in 143 of 143 readingsLast checked 2026-09-13 15:50 UTCWeb app
OperationalThe application that serves every page and every action you take.
Up in 143 of 143 readingsLast checked 2026-09-13 15:50 UTCSecure front door
OperationalThe encrypted entry point your browser connects to.
Up in 143 of 143 readingsLast checked 2026-09-13 15:50 UTCScheduled jobs
OperationalReminders, digests, calendar sync and everything else that runs on a clock.
Up in 143 of 143 readingsLast checked 2026-09-13 15:50 UTCCall media relay
OperationalCarries audio and video when two people on a call cannot reach each other directly.
Up in 143 of 143 readingsLast checked 2026-09-13 15:50 UTCOutbound email
OperationalInvitations, notifications and password resets leaving the platform.
Up in 143 of 143 readingsLast checked 2026-09-13 15:50 UTCMeeting recorder
OperationalThe assistant that joins a meeting to record and transcribe it.
Up in 143 of 143 readingsLast checked 2026-09-13 15:50 UTCAutomation engine
OperationalRuns the workflow rules a workspace has set up.
Up in 143 of 143 readingsLast checked 2026-09-13 15:50 UTCInternal database consoleInternal tool
OperationalA tool our own engineers use. No page, job or request in Noots depends on it.
Up in 143 of 143 readingsLast checked 2026-09-13 15:50 UTC
Recently resolved
Everything we closed in the last 30 days, with the timeline exactly as it was written at the time.
Nothing has been opened and closed in that window.
What we measure, and what we will not claim
A scheduled job asks each of the services above a direct question every ten minutes, and we keep the answer for fourteen days. Each bar on a service is one 30 minute window and it shows the WORST answer inside that window, so a service that failed for ninety seconds and recovered still paints that window: the point of keeping history is to see the blip nobody was watching for.
A faded bar means nobody checked in that window. That is a different fact from a service being fine, and we will not draw it the same way.
We publish a count of readings, not an uptime percentage. A check that did not run records nothing at all, so any percentage computed from these numbers would get BETTER during an outage that also stopped the checks. Counting the readings is the honest version of the same information, and it tells you how much evidence is behind each row.
The list above is every service we probe, including one that only our own engineers use. That one is labelled, and it is left out of the headline: an internal tool being restarted is not a reason to tell you your work is at risk.
Incidents on this page are written by a person, not generated. Our checks cannot see everything, so a green board and an open incident are not a contradiction.