GitHub Availability Report: March 2023

In March, GitHub experienced six incidents causing degraded performance across its services, with a notable incident from February 28 affecting GitHub Codespaces in the East US region due to slow VM allocation from a cloud provider. This incident prompted a redirection of codespace creations and led to architectural changes for quicker failover. On March 1, latency issues in package registries were traced to an unhealthy disk on a VM node, prompting database failover and a shift to a new MySQL infrastructure for improved reliability. March 2 saw GitHub Actions workflow failures due to an SSL certificate issue with a CDN provider, which was resolved by removing the incorrect binding. On March 15, increased latency in package registries resulted from a slow-running query during maintenance, leading to updated safety checks in the maintenance process. Incidents on March 27, 29, and 31 affected pages, codespaces, and Git operations, with investigations still ongoing. GitHub is implementing changes to improve monitoring, failover capabilities, and infrastructure reliability, with updates and ongoing efforts shared on their status page and engineering blog.