GitHub outage history

GitHub Actions ran into degraded performance starting at 3:21 PM UTC, as an upstream provider began throttling requests and returning 429 responses, leading to delays in starting Actions jobs. Users reported slow runs and occasional unresponsiveness as the issue unfolded. By 4:11 PM UTC, GitHub cautioned that customers could experience delays of up to ten minutes while the situation was investigated.

A mitigation was put in place with an expected recovery window of about thirty minutes by 4:51 PM UTC. At 5:51 PM UTC, GitHub noted that degraded performance persisted due to throttling within an upstream Azure dependency, but capacity had recovered and the team continued to monitor while working with Azure on the root cause. The service returned to normal by 6:03 PM UTC.

On 2:51 PM UTC, GitHub reported degraded performance for Actions as run-start delays affected Ubuntu runners, causing intermittent delays starting workflows and longer CI times for users. The issue advanced as GitHub identified the cause and deployed a fix; by 3:06 PM UTC Ubuntu run-start delays were resolved while Windows runner delays remained under investigation, and the Actions component returned to normal by 3:11 PM UTC.

Users experienced elevated latency affecting GitHub web requests, with pages loading slowly and some API calls taking longer than usual. The issue began at 1:42 PM UTC when GitHub flagged a latency concern and continued to impact users through about 1:56 PM UTC. Report activity started earlier, with user submissions beginning at 1:00 PM UTC and peaking around 1:15 PM UTC.

GitHub's status updates indicated an investigation into the elevated latency affecting web requests, noting ongoing work and promising updates as information became available. After a brief window, the service returned to normal, with the status page indicating a full recovery by 1:56 PM UTC.

Users attempting to use GitHub Copilot Code Review could not complete reviews, disrupting workflows for some teams. The issue began with degraded performance at 9:21 PM UTC and escalated to a full outage for Copilot Code Review at 9:26 PM UTC as GitHub rolled back the impacted change. The outage persisted for roughly 45 minutes and was resolved when services returned to normal at 10:11 PM UTC; GitHub noted that the revert was in flight and full recovery was expected once deployment completed.

An issue with GitHub's Billing information updates disrupted creating or updating billing details for some users. The disruption began at Sep 24, 2026 4:56 PM UTC and affected the Billing information updates component, with users experiencing errors when attempting billing changes. A lone user report described a 'not found' error during billing-related tasks.

GitHub's status page documented a mitigation process: the problem was identified and mitigation efforts started, with updates at 5:46 PM UTC and 6:26 PM UTC. By 8:11 PM UTC the team said mitigation had been applied and signs of recovery were visible, and the service returned to normal by 8:26 PM UTC. The incident prompted ongoing monitoring to ensure stability.

From around 10:16 AM UTC, developers began experiencing degraded performance on GitHub's API, with the API Requests component listed as degraded. By 10:21 AM UTC, GitHub stated that database replicas had detached and that teams were working to restore them, warning that users could see issues beyond creating organizations and a degraded experience with the GitHub API and Projects. The impact was global for API usage and project-related features, affecting workflows that rely on API calls, project creation, and automation.

GitHub experienced degraded performance affecting the Pull Requests workflow. Users attempting to open, review, or merge pull requests likely faced slower responses and intermittent delays while the issue was active. The GitHub status page indicated an incident with Pull Requests and that the team was investigating reports of degraded performance, with updates showing the Pull Requests component moving from up to warn and then back to up as the issue was resolved.

GitHub Copilot relied on OpenAI models that power Copilot AI Model Providers experienced an elevated rate of errors, resulting in degraded performance for Copilot features. The incident began at 9:01 PM UTC and was resolved by 9:41 PM UTC, according to GitHub's status updates, which documented the elevated errors and that teams were investigating. After remediation, normal Copilot functionality resumed. The disruption primarily affected Copilot components that depend on these OpenAI model providers, with users potentially seeing errors or slower responses during the window of impact, and the issue lasting about forty minutes.

GitHub users experienced degraded Copilot availability as Gemini 3.8 Flash issues affected Copilot AI Model Providers. The incident began around 7:21 AM UTC, with Copilot AI Model Providers entering a degraded state and the status page noting an ongoing investigation. The situation persisted, and a second update at 8:31 AM UTC confirmed degraded performance as the team continued to work on a fix. A single user report around 11:45 AM UTC described being unable to use Copilot, illustrating the impact on developer workflows. The problem was resolved by 11:46 AM UTC, with Copilot AI Model Providers returning to an up state and normal service gradually restoring.

Some GitHub services experienced disruption, resulting in degraded performance for a portion of users. The incident began around 7:16 PM UTC when GitHub announced that they were investigating reports of impacted performance across certain services.

By 7:56 PM UTC the issue was resolved and normal service resumed, with GitHub returning to typical performance. The disruption lasted around forty minutes and was limited to a subset of services.