Anthropic had two separate incidents on 16 November. The first, elevated errors across Claude Sonnet 5.5 and other Claude models, lasted 1h 18m and was resolved by Anthropic. The second, a fault stopping Claude from joining new calls in Claude Meetings (beta), was still open at 18:10 UTC, the last update Anthropic published.
Started
07:49 UTC
Duration
Lasted 1h 18m
Source
IsDown
Next time
Anthropic 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.
Affected components: claude.ai, api.anthropic.com
What happened?
Anthropic's status page logged two distinct problems on 16 November. The first began at 07:49 UTC as elevated errors affecting Claude Sonnet 5.5 and Claude Fable 5.2. Anthropic deployed a fix by 08:02 UTC and marked the issue resolved at 09:04 UTC, stating the impact actually ran from 23:32 PT on 15 November (07:32 UTC on the 16th) through 00:45 PT (08:45 UTC), covering claude.ai, the Claude API, Claude Code and Claude Cowork. That gives a duration of 1h 18m, in line with Anthropic's 90 day median incident length of 1 hour 16 minutes, so this was a typical-length event for them rather than an outlier. The second incident started separately at 15:10 UTC and concerned Claude Meetings (beta), with Claude unable to join new calls. Anthropic said calls already in progress were unaffected and that core functionality was not impacted. By 16:27 UTC they had paused Auto-join entirely while investigating, and by 18:10 UTC they confirmed Claude might still be unable to join some calls, with Auto-join still paused and no fix shown. Anthropic explicitly said they expected this one to take some time to fully resolve. No cause has been published for either incident. No user reports were filed during either window. Two components were affected across the day: claude.ai and api.anthropic.com. Anthropic's own 90 day record shows 46 incidents, so a two-incident day sits within their normal pattern rather than against it.
Learning
This is a case where the failure sat inside a specific feature path rather than across the whole API, which is exactly the kind of fault that a simple uptime check on your own endpoint will not catch. The first incident was a model-serving error rate problem that Anthropic fixed within the hour. The second was narrower still: new calls failing to join a beta meetings feature while calls already running kept working, which means any monitor only watching that core API would have stayed green throughout, correctly, because that core function genuinely held up. Your own monitoring tells you whether your product is reachable and responding, but it cannot tell you whether the fault sits upstream at your model provider versus in your own integration code. Anthropic's 30 day figures show 19 incidents with 17 resolved and a median length matching the November 16 morning event almost exactly, so these interruptions are a regular feature of running on their platform, not a one-off. When a feature like Claude Meetings breaks while the main API stays healthy, you need visibility into that specific vendor-side path, not just a ping to a root domain. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.