Heroku's Node.js build pipeline broke for 5h 14m on 28 October after an upstream dependency started throwing elevated error rates. If you tried to deploy a Node.js app on Heroku during that window, the fault was on Heroku's side, not yours.
Started
10:55 UTC
Duration
Lasted 5h 14m
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: Tools
What happened?
Heroku logged one incident on 28 October 2026, titled "Heroku Feature Degradation", affecting the Tools component. It started at 10:55 UTC and ran for 5h 14m before resolution. The first update, at 10:57 UTC, pinned the problem on an upstream dependency causing elevated error rates specifically in Node.js builds. Heroku posted four updates in total across the window. By 15:02 UTC the team said most upstream errors had been triaged and told users to keep retrying failed builds while they worked toward full restoration. A near-identical update followed at 15:31 UTC, again asking users to retry and promising a further update once normal operations were fully restored. No user reports were filed during the window on this page's record. Heroku's status page does not explain what the upstream dependency was or why it failed. Over the prior 90 days Heroku logged 4 incidents with a median duration of 1 hour 24 minutes, so this one ran well past the company's own typical recovery time. Across all time Heroku has logged 378 incidents, averaging 4.8 a month.
Learning
This incident sat at the build layer, inside Heroku's own pipeline for compiling Node.js packages, which is not something an external uptime check against your app's URL would ever see. Your monitoring watches whether your live product responds, and it can stay green the whole time because a build failure does not take your already-running app offline, it only blocks new deploys. That is why UptimeRobot's own 30 day stats for this provider show one incident, resolved, with a median duration of 5h 14m, matching this outage exactly rather than conflicting with a clean uptime graph on your end. The two readings are not contradictory: one tells you your product was up, the other tells you Heroku's build infrastructure was not. If you had a deploy queued during this window, retries were the only workaround Heroku offered, and that advice stood for over five hours. Watching your own endpoints tells you about your half of the stack. UptimeRobot now watches the provider's half too, so next time you know which side broke without guessing.