If your macOS builds on CircleCI crawled on 20 October 2026, it was not your pipeline. CircleCI's m4pro.large resource class saw queued jobs wait up to 30 minutes at the peak, for 2h 16m, a fault on CircleCI's side of the build infrastructure.
Started
15:58 UTC
Duration
Lasted 2h 16m
Source
IsDown
Next time
CircleCI 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: macOS Jobs
What happened?
CircleCI logged one incident on 20 October 2026, affecting macOS Jobs, starting at 15:58 UTC and running for 2h 16m until resolution at 18:14 UTC. The title CircleCI gave it was "Increased wait times for macOS jobs on m4pro.large", and that is exactly what the five updates describe throughout. At 15:58 CircleCI said it was seeing increased wait times on the m4pro.large resource class and was investigating. By 16:32 it confirmed jobs were waiting up to 30 minutes to start. At 17:02 it attributed the delay to heavy demand on the large class specifically, noting m4pro.medium was unaffected throughout. The 17:31 update gave the clearest picture of scale: Xcode 27.0 jobs had peaked at roughly 40 minutes of wait and were down to about 3.5 minutes, while jobs on Xcode 26.6 and 26.4.1 were still clearing a backlog. CircleCI closed the incident at 18:14, reporting wait times back to normal. No user reports were filed during the window on this record. CircleCI's status page does not explain why demand spiked at that moment, only that it did and that it was confined to the large resource class. This was CircleCI's own account of its own infrastructure, so it carries the weight of a single vendor filing, not an independent confirmation.
Learning
This incident sat entirely inside CircleCI's build queue, a layer that an external uptime check against CircleCI's public endpoints would not touch, because the API and dashboard can answer normally while jobs queue behind a resource class that is short on capacity. That is the general shape of this kind of failure: nothing is down, things are just slow, and slow does not trip a binary up or down check on either side. Over the past 90 days CircleCI has logged 26 incidents with a median duration of 1 hour 50 minutes, and this one ran close to that at 2h 16m, so queuing delays on specific resource classes are a recurring pattern rather than a one off. If your own monitoring only watches your application's endpoints, it has no visibility into CircleCI's job scheduler, and it will stay green while your builds pile up. Watching your own product's health and watching the provider your builds depend on are two different jobs. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.