GitHub Actions was down yet again

GitHub Actions encountered another disruption on Wednesday, just days after the code host reassured users of improved reliability. The issue, as previously reported, originated at 1511 UTC when GitHub identified a problem with a database primary and failed over to a replica. However, the move did not fully address the degradation.

In response, GitHub throttled inbound traffic while investigating upstream Vitess issues before gradually restoring traffic. By 1800 UTC, the platform was reported to be functioning as expected with recovered inbound queues.

This latest disruption adds to a growing list of issues GitHub has faced this year, with the platform experiencing a total of 26 problems in April alone. May, June, July, and August have all seen 23 incidents each, with August nearing its end and just under a week left. The year-to-date reliability issues have amounted to at least 23 incidents per month. GitHub has struggled to maintain triple nines uptime, currently showing 98.13 percent as of Wednesday.

The platform's central role in many developers' CI/CD workflows has made it particularly vulnerable to GitHub's ongoing reliability challenges. August has been particularly troublesome for GitHub, with a nearly eight-hour outage on the 17th affecting multiple services, including Issues, Pull Requests, APIs, Actions, and Copilot.

GitHub has attributed many of its issues to AI, attributing the August 17 outage to a surge in usage from bots and agents that the company had difficulty managing. GitHub's CTO, Vladimir Fedorov, has apologized for the incident and promised to scale the platform to better support its growing user base. However, even with Fedorov's assurances, GitHub's reliability continues to slip.