Notion's MCP integration broke for 1h 22m on 22 October 2026, specifically the query_data_source tool used by Claude clients. The fault sat on Notion's side, not yours.
Started
03:23 UTC
Duration
Lasted 1h 22m
Source
IsDown
Next time
Notion 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 03:23 UTC when Notion opened a single report: the MCP query_data_source tool was temporarily unavailable in Claude clients. That is the only update Notion filed for this incident. There is no follow-up confirming a fix, no root cause note, and no explicit resolution message in the record, though the day's total duration is given as 1h 22m. No components are listed as affected on Notion's status page for this entry, and no user reports were logged in this window. Notion's own wording frames this narrowly: a specific tool, inside a specific integration path, for a specific client. That is a much smaller claim than "Notion is down", and the record does not support stretching it any further. Over the past 90 days Notion has logged 31 incidents with a median duration of 2 hours 28 minutes, so this one resolved faster than typical. Across its full history, Notion has averaged 4.3 incidents a month over 327 total incidents, putting outages or partial faults as a regular feature of running on this provider rather than a rare event. The single filed update means you are relying almost entirely on Notion's own account of what broke and for how long, since no second source corroborates it here.
Learning
This incident was narrow by design: a named tool inside an integration path, not the whole product, and Notion's status page only ever described it that way. If your own monitoring targets Notion's main app or API and not this specific MCP tool path, it would have stayed green the entire time, and that green reading would have been accurate for what it was checking. That is the general shape of these failures: a single downstream tool or integration can fail while the core service keeps answering everywhere else, so a broad health check simply never sees it. UptimeRobot's own 30 day data shows 12 incidents against Notion with a median duration of 2h 55m, which is longer than this one, suggesting this failure cleared quicker than Notion's recent norm. Watching only your own endpoints tells you your product is up, not whether the specific provider feature you depend on is working. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.
