Another GitHub outage takes down Actions and Pages, the sixth incident this month

GitHub Actions and GitHub Pages went down starting at 15:22 UTC on August 6, and the incident was still unresolved two and a half hours later, per GitHub's own status updates. Workflow runs failed to start or crashed mid-execution, and users hit unexpected rate limits and API errors. Copilot code review, Copilot's coding agent, hosted runners, GitHub Enterprise Importer migrations, and webhook deliveries were all affected alongside the two headline services. GitHub said it was rolling out a further fix but didn't specify a root cause.

This wasn't an isolated event. GitHub's own status history shows a similar pattern in the months before this one: 26 incidents in April, 23 in May, 23 in June, 26 in July, and 6 more in the first six days of August alone, with a similar Actions outage on July 29, just over a week before this one. Terraform creator Mitchell Hashimoto has said publicly the platform is no longer reliable enough to host serious work.

For a team running CI/CD entirely on GitHub Actions, the practical takeaway isn't to panic, a two-job test-and-deploy pipeline has no real self-hosted alternative that costs less operational overhead. It's to know what actually breaks when GitHub does: deploys stall, not code. Nothing in a minimal pipeline is destructive on failure, it just doesn't run, which is exactly the tradeoff worth understanding before the next outage lands mid-deploy.