ChatGPT Work saw degraded performance for 27 minutes on 28 October 2026. OpenAI flagged the issue itself, and the disruption was on their side, not yours.
Started
03:39 UTC
Duration
Lasted 27m
Source
IsDown
Next time
OpenAI 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?
OpenAI's status page logged one incident on 28 October 2026, titled "Degraded Performace on ChatGPT Work", starting at 03:39 UTC and lasting 27 minutes. The only update on record is the opening one: OpenAI said they were investigating a potential issue for the listed services and checking how it affects users, and that they would share more when they knew it. No further update was published after that, and no closing note appears in the record, though the duration shown is 27 minutes. OpenAI did not list which components were affected beyond naming ChatGPT Work in the title, so there is no breakdown of what exactly degraded, whether response latency, completion quality, or something else. No user reports were filed during the window, so there is no outside account of what the degradation felt like in practice. This is a single filing from OpenAI, and on its own it tells you that something was off for 27 minutes and not much more. Over the past 90 days OpenAI has logged 103 incidents with a median duration of 1 hour 39 minutes, and across all time they average 15.4 incidents a month. Measured against that median, 27 minutes is a short one. The cause is not on the record, OpenAI's status page does not say why performance degraded.
Learning
This incident was short and resolved quickly, but it sits inside a pattern of frequent degradations from OpenAI, and short incidents are often the hardest ones for your own monitoring to catch. If you are only polling your own endpoints or watching your own error rates, a 27 minute dip in a provider's performance can come and go without tripping any alert on your side, because your systems may simply retry, queue, or serve slightly slower without failing outright. That is the gap between watching your own half of the stack and watching the provider's half. UptimeRobot's 30 day figures for OpenAI show 45 incidents, all resolved, with a median duration of 2 hours 7 minutes, which is a different measurement angle from OpenAI's own 90 day median of 1 hour 39 minutes, but both point the same way: degradations are common and often brief enough to miss. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.