Microsoft 365 Copilot Chat stopped returning responses to some queries for 3h 58m on 11 November 2026. Microsoft traced the fault to its own backend infrastructure, not your account or network.
Started
14:39 UTC
Duration
Lasted 3h 58m
Source
IsDown
Next time
Microsoft 365 is degraded. Is your product working?
An incident on their side does not always mean an outage on yours, and the only way to know is to already be watching your own endpoints. Set that up in under a minute and you will be notified the next time your own site is affected. And there will be a next time.
50 monitors for your own websites. No card.
What happened?
The incident started at 14:39 UTC and ran for 3h 58m before Microsoft closed it. Microsoft's own write-up describes the user-facing symptom as queries submitted to Copilot Chat getting no response at all, sometimes surfacing as "Something went wrong" or "I couldn't help with that" errors. The first update, at 14:43 UTC, said Microsoft was still reviewing service monitoring telemetry to find a root cause. By 16:52 UTC Microsoft had narrowed it to a rise in "499" error codes from the portion of its infrastructure that handles Copilot Chat query responses, and said that infrastructure wasn't processing request traffic at normal efficiency. Microsoft also floated a second theory, that request traffic may have been routed incorrectly, without settling on either explanation. The same update scoped the impact to some users in West Europe, a narrower footprint than the opening description suggested. Microsoft reported the error rate was already falling and called it a self-recovering situation while it kept watching its own telemetry. No components list was published alongside this incident. There were no user reports filed in this window, and no further update followed to confirm the fix or give a final root cause. Two updates is what Microsoft put on record for a fault lasting nearly four hours.
Learning
This incident sat entirely inside Microsoft's backend, specifically the infrastructure that handles Copilot Chat query responses, which is not something an external check against your own endpoints would ever see. Your monitoring tells you your product is up and taking requests, and it was right to say so, because the fault was downstream in Microsoft's systems, not in anything you run. Microsoft's own account shows the shape of this kind of failure: errors creep up at the infrastructure layer, get scoped only once engineers dig into telemetry, and can resolve on their own before anyone fully explains why. Over the past 90 days Microsoft has logged 62 incidents on this service with a median duration of 9 hours 3 minutes, so a fault lasting under four hours is on the shorter side for them. If a Copilot-dependent feature in your product goes quiet, this is the kind of upstream cause worth checking first. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.