GitHub Actions Outage Spans Over Eight Hours, Cascading Into Pages, Copilot, and Webhook Delivery
A GitHub Actions failure on August 6 knocked out workflows, Pages, and Copilot for over eight hours, extending a run of dozens of monthly incidents this year.
Overview
GitHub Actions, the platform’s CI/CD automation service, failed for hours on August 6, with GitHub first reporting “degraded performance” at 15:22 UTC before the incident spread into GitHub Pages, Copilot code review, Copilot coding agent, hosted runners, and webhook deliveries, according to The Register. GitHub’s own status page shows the incident was not marked resolved until 02:04 UTC the following morning, according to GitHub Status.
What We Know
GitHub reported Actions suffering “degraded performance” at 1522 UTC on August 6, escalating to “degraded availability” about 20 minutes later, according to The Register. Workflow runs began failing to start or failing partway through execution, requests to the Actions REST API returned errors, and users saw unexpected rate limits, the outlet reported.
By 1740 UTC, GitHub said in a status update that “Engineers have applied a number of mitigations and are rolling out a further fix across all affected systems now,” adding that “Copilot code review, Copilot coding agent, hosted runners, and migrations using GitHub Enterprise Importer” were also affected, and that webhook deliveries might be delayed, according to The Register.
GitHub’s own status history shows engineers identifying the mechanism behind the failure by 23:13 UTC, posting: “We have deployed fixes that address runners being assigned invalid jobs and are taking additional steps to clear the backlog.” At 00:01 UTC on August 7, GitHub reported that “System-wide queues have been drained, and new jobs are being processed as expected,” and by 00:06 UTC the degradation affecting Actions and Pages was marked as mitigated, according to GitHub Status.
A residual problem lingered afterward: GitHub’s status page noted at 00:59 UTC that some Actions Runner Controller runners were “taking longer than expected to recover,” and a follow-up update at 02:03 UTC explained that some Actions Runner Controller runner pods had become “stuck in an idle state” during the incident. GitHub marked the incident fully resolved at 02:04 UTC on August 7, according to GitHub Status.
The August 6 incident extends a run of frequent GitHub disruptions. The Register counted 26 incidents logged on GitHub’s status page in April, 23 in May, 23 in June, and 26 in July, with a similar Actions outage occurring just days earlier, on July 29, according to The Register. GitHub has pointed to AI-driven load as a factor in its reliability struggles; in its May 2026 availability report, GitHub senior vice president Jakub Oleksy wrote that “We are making structural changes that permanently remove failure modes,” according to The GitHub Blog.
The recurring outages have already driven at least one prominent departure. In an April 28 post explaining his decision to move the Ghostty terminal emulator project off GitHub, developer Mitchell Hashimoto wrote that after keeping a month-long journal of outages that disrupted his work, “Almost every day has an X,” and that GitHub “is no longer a place for serious work if it just blocks you out for hours per day, every day,” according to Mitchell Hashimoto. That post predates the August 6 incident and was written in response to the broader pattern of outages rather than this specific one.
What We Don’t Know
GitHub has not published a detailed root-cause explanation beyond the brief status-page description of runners being “assigned invalid jobs,” and the company did not respond to The Register’s request for additional information on the incident, according to The Register. It is not yet clear when GitHub Enterprise Importer migrations, which the company said were affected alongside Actions, fully returned to normal.