GitHub’s continuous integration and delivery platform, Actions, experienced another outage on Wednesday, August 26, further denting confidence in the service after a series of reliability issues throughout 2026. According to GitHub’s incident report, problems began at 1511 UTC when a database primary failed. A failover to a replica did not fully resolve the degradation, prompting GitHub to throttle inbound traffic while investigating upstream Vitess issues. By 1800 UTC, GitHub declared Actions operational and inbound queues recovered.
The outage comes just six days after CTO Vladimir Fedorov published a postmortem for a nearly eight-hour outage on August 17 that hit multiple services including Actions, Issues, Pull Requests, and Copilot. Fedorov admitted the platform had let users down and pledged to scale infrastructure to support a growing user base—human and automated alike. “We’ll earn your trust through the scaling and reliability of the platform,” he wrote. Wednesday’s disruption suggests those efforts have yet to stabilize the service.
GitHub’s reliability record this year has been persistently poor. The company’s status history shows 26 incidents in April, 23 in May, 26 in June, 26 in July, and at least 23 so far in August with days remaining. February set the 2026 high with 37 incidents; March logged 32. GitHub’s own uptime page for Actions currently stands at 98.13% for August—dangerously close to the 97% threshold that would mark a serious degradation for a mission-critical SaaS product.
GitHub attributes much of the strain to AI-driven usage. The company claims it now handles 2.9 billion commits, 24 million new repositories, and 130 million merged pull requests per month—roughly double the commit volume it processed in April. Bots and agents, Fedorov has stated, are generating unprecedented traffic that the platform was not designed to accommodate. GitHub has been investing in scaling, but the frequency of outages suggests that demand is outpacing upgrades.
The Register contacted GitHub for comment on the latest outage but received no response. For many developers, Actions is an indispensable part of their workflow. Each disruption forces teams to delay builds, troubleshoot failures, and reconsider reliance on a single vendor. While GitHub’s scale is unmatched, competitors including GitLab and self-hosted runners offer alternatives that may appeal to organizations seeking greater reliability.
As August draws to a close, GitHub faces a critical test: can it finally deliver on its promises before September adds another chapter to a year of discontent?