Heroku's Private Spaces in the Tokyo region had trouble booting new dynos for just over three hours on 11 November 2026. Heroku traced the fault to an upstream cloud provider, not to its own platform.
Started
18:18 UTC
Duration
Lasted 3h 8m
Source
IsDown
Next time
Heroku 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: Apps
What happened?
The incident started at 18:18 UTC and ran for 3h 8m before resolution. Heroku's status updates describe a service degradation affecting Private Spaces in the Tokyo region, caused by an issue with an upstream cloud provider. The practical symptom was that customers could not restart dynos or scale up, because new instances could not boot. Heroku was explicit that running dynos were not expected to be impacted, so apps already up stayed up, but anything needing a fresh instance was stuck. Three updates were posted, at 18:16, 19:17 and 20:18 UTC, each repeating that the upstream provider was reporting progress but that some customers might still see failures when restarting or scaling. No update names the upstream provider or explains what actually went wrong on their end. Only one component is listed as affected: Apps. No user reports were filed during the window, which is consistent with a regional, dyno-lifecycle-specific fault rather than a broad platform failure. Over the past 90 days Heroku has logged 4 incidents with a median duration of 2 hours 22 minutes, and this one ran longer than that median. Across all time Heroku averages close to 5 incidents a month, so a single-region degradation like this sits within its normal pattern rather than standing out as unusual.
Learning
Heroku's own write-up puts the fault with an upstream cloud provider, which means the failure sat below the platform layer you can see or control, in the infrastructure that boots new dyno instances. A check against your app's URL would likely have stayed green the whole time, because running dynos were not affected, only the ability to create new ones. That is the trap with scaling or restart failures: your endpoint looks fine until the moment you actually need capacity, and by then the gap is already costing you. UptimeRobot's own 30 day figures show 2 resolved incidents on Heroku with a median duration of 4h 11m, longer than Heroku's self-reported history, which is the kind of discrepancy that shows up when you watch the provider's status directly rather than relying on your app's uptime alone. Your monitoring covers your half, UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.