GitHub said Actions workflows were failing and later added Pages and other developer services to the outage.
On 6 August, GitHub’s status updates showed problems first in Actions, the platform’s automation service for running software workflows, and later in Pages, its hosting service for websites built from repositories. The Register reported that GitHub first logged degraded performance at 1522 UTC, then said availability was also degraded around 20 minutes later.
For users, this was not an abstract infrastructure issue. GitHub said workflow runs were failing to start or were stopping partway through execution, and that requests to the Actions REST API were returning errors. That means teams using GitHub to automatically test code, build software or deploy updates could find those routine steps blocked or unreliable.
What failed
According to The Register’s account of GitHub’s status page, the incident started with Actions. GitHub first described the problem as degraded performance, then widened that to degraded availability. A few minutes later, the company said workflow runs were either not starting at all or were failing during execution.
GitHub also said requests to the Actions REST API were returning errors and that users were seeing unexpected rate limits. In simpler terms, both the background jobs and the programming interface that tools use to interact with those jobs were affected. Shortly after GitHub said it was working to mitigate the Actions problem, Pages was added to the list of affected systems.
The source also says GitHub identified a cause, but had not named it publicly at that point. As of GitHub’s most recent update cited by the report, engineers had applied several mitigations and were rolling out a further fix across affected systems, but the issue had not yet been resolved.
Which services were affected
The disruption was not limited to Actions and Pages. The same report says GitHub’s status updates later listed Copilot code review, Copilot coding agent, hosted runners, and migrations using GitHub Enterprise Importer as affected too. Webhook deliveries might also be delayed.
Those names describe different parts of GitHub’s developer platform. Hosted runners are GitHub-managed machines that execute automated jobs. Webhooks are automated notifications sent to other systems when something changes in a repository. Enterprise Importer is used for migrations. If these are delayed or unavailable, the knock-on effect can spread beyond a single failed build to other connected systems and team processes.
Pages matters in a different way. It is GitHub’s service for publishing websites from repository content, often used for project pages and documentation. The report describes Pages availability as degraded, which suggests some users may have had trouble accessing or publishing those sites, though the source does not quantify the extent.
This follows a string of GitHub incidents
The Register places the outage in the context of repeated GitHub stability issues. It says a similar Actions outage happened on 29 July. It also cites GitHub’s status history page as listing 26 incidents in July, 23 in June, 23 in May and 26 in April. Six days into August, the status page listed six incidents for the month so far, according to the report.
Incidents cited from GitHub’s status history in the source report
The source further says GitHub apologised in April for continued outages and in June promised structural improvements. It attributes to GitHub SVP Jakub Oleksy a claim that those changes would “permanently remove failure modes”. However, the same report argues that the status history since then does not yet show a clear improvement. That is an interpretation by the publication, not a statement from GitHub that the problem has been solved or not solved.
Why the broader reliability question remains open
The strongest conclusion supported by the source is narrower than a claim that developers are abandoning GitHub. This outage shows how problems in one managed developer platform can ripple through builds, deployments, hosted pages and connected automation at the same time.
The report says some developers have considered moving projects away from GitHub because of repeated stability issues, and it cites Ghostty developer Mitchell Hashimoto’s criticism of the platform’s reliability. But the same source does not establish a broad user exodus, how many teams are actually leaving, or whether most organisations are changing their tooling.
The clearer implication is operational. GitHub said it had applied mitigations and was rolling out another fix, yet the incident counts cited in the report suggest reliability remains a live concern for teams that depend on Actions, Pages and related managed services. For IT and engineering leaders, the narrow takeaway is not that a market shift has already happened, but that repeated outages in a core developer platform can expose how much of a delivery pipeline relies on a single provider.