GitHub says its Azure migration reduced load but is still incomplete
GitHub made progress on moving production traffic and tightening failure handling, but recent incidents and a missed migration target show the work is not done.
GitHub said its August and September work reduced pressure on shared databases, added overload controls, improved failure containment for pull requests, and shifted more production traffic to Azure, but it also missed a September 30 migration target and had recent incidents that showed the effort is still unfinished. By October 7, 120 of 160 production services were running in Central US, 13 of 51 primary MySQL databases had moved, and Azure was carrying major portions of several workloads, including GitHub’s core app, Git read traffic, and stream-processing traffic. GitHub paused Central US deployments on September 11 because capacity was uneven across availability zones, then resumed ramps on September 18 and said it still aims to unlock all available capacity by the end of October. The report also says a background cleanup job drove a shared database to its connection limit on September 13, while a September 28 Copilot code review service change caused some customers to receive only a limited result and left reviews incomplete even though the job reported success.
Why it matters
For customers, the report points to a service that is becoming more resilient in some areas while still exposing users to partial outages and incomplete automated reviews. The missed migration target and the pause in deployments also show the transition is still constrained by capacity and operational risk, so the remaining shift to Azure matters less as an internal milestone than as a sign of how much stability GitHub still has to prove.
Keep or strike?
Does this story matter, or is it hype? Mark it before you see what everyone else did.
Sources
- GitHub Blog