GitHub Actions was down yet again

On Wednesday, GitHub Actions experienced another outage, following a series of reliability issues that have plagued the code host's CI/CD platform. According to the incident report, the problem began at 1511 UTC when the source code host identified an issue with a database primary and failed over to a replica. However, this move did not fully resolve the degradation, leading GitHub to throttle inbound traffic while investigating upstream Vitess issues. By 1800 UTC, the platform was reported to be operating as expected, with inbound queues recovering.

The latest disruption is particularly concerning given GitHub's claims of improved reliability. Last week, GitHub stated that it was now serving double the commits it was handling in April, a month that saw many reliability issues. This year, GitHub has suffered at least 23 reliability issues each month, with August 2023 being particularly bad so far.

The platform has faced numerous outages, including a nearly eight-hour outage on August 17 that affected multiple services, including Actions, Issues, Pull Requests, APIs, and Copilot.

GitHub's CTO, Vladimir Fedorov, acknowledged the company's shortcomings in handling the increased usage caused by AI bots and agents. He promised to scale and improve the platform's reliability in a postmortem of the August 17 outage. However, six days later, the reliability issues persist, with the number of incidents growing. Despite GitHub's claims of handling 2.9 billion commits, 24 million new repositories, and 130 million merged pull requests monthly, the uptime for Actions is currently at 98.13 percent, putting the platform at risk of slipping into a 97 percent reliability territory.