Disruption with Copilot Cloud services
Facts
- Source
- GitHub Status · original
- Started
- (15 days ago)
- Ended
- · lasted 2 h 30 min
- How it ended
- Microsoft marked it resolved.
- First captured by us
- Last changed
- (43 min ago)
Last text from Microsoft
On October 29, 2026, between 05:50 UTC and 12:40 UTC, Copilot agent sessions on GitHub.com and in VS Code were degraded. Live session updates often failed to load, and on average 2.14% of requests to the agent sessions service failed, with a peak of 8.61%. For about 10 minutes around 09:20 UTC, up to 40% of requests to list or create agent sessions failed. Copilot Chat on GitHub.com showed degraded agent session features during this time. The issue occurred when a large increase in task-list requests from a client experiment overloaded the cache that supports the agent sessions service. We mitigated the impact by increasing cache capacity, reducing nonessential load on the cache, and disabling the client experiment. We are moving rate limiting to separate infrastructure, reducing client polling, and improving our alerting so we can detect and respond to this type of overload faster.
Quoted from GitHub Status (source), captured . Reproduced for reporting and archival purposes.
Official updates
Posted by GitHub Status, newest first.
Resolved
On October 29, 2026, between 05:50 UTC and 12:40 UTC, Copilot agent sessions on GitHub.com and in VS Code were degraded. Live session updates often failed to load, and on average 2.14% of requests to the agent sessions service failed, with a peak of 8.61%. For about 10 minutes around 09:20 UTC, up to 40% of requests to list or create agent sessions failed. Copilot Chat on GitHub.com showed degraded agent session features during this time. The issue occurred when a large increase in task-list requests from a client experiment overloaded the cache that supports the agent sessions service. We mitigated the impact by increasing cache capacity, reducing nonessential load on the cache, and disabling the client experiment. We are moving rate limiting to separate infrastructure, reducing client polling, and improving our alerting so we can detect and respond to this type of overload faster.
Monitoring
The degradation has been mitigated. We are monitoring to ensure stability.
Investigating
We continue to experience failure rates of 1-2% and increased latency on our agent sessions service, including creating and listing sessions. This can impact cloud based agent sessions on web, mobile, VS Code, CLI and the GitHub Copilot app. We are actively applying mitigations and monitoring for recovery.
Investigating
We continue to experience failure rates of 1-2% and increased latency on our agent sessions service, including creating and listing sessions. This can impact cloud based agent sessions on web, mobile, VS Code, CLI and the GitHub Copilot app.
Investigating
We are experiencing increased failures and latency on our agent sessions service, including creating and listing sessions. This can impact cloud based agent sessions on web, mobile, VS Code, CLI and the GitHub Copilot app.
Investigating
We are investigating reports of impacted performance for some GitHub services.
What Microsoft changed
Microsoft edits the text of an incident in place. We keep every version we capture (checked every minute) and highlight the difference. Removed text is struck through, added text is marked.
First version we captured
Captured · Microsoft's own timestamp · revisionz3gdmb1hOriginal text as we first saw it
On October 29, 2026, between 05:50 UTC and 12:40 UTC, Copilot agent sessions on GitHub.com and in VS Code were degraded. Live session updates often failed to load, and on average 2.14% of requests to the agent sessions service failed, with a peak of 8.61%. For about 10 minutes around 09:20 UTC, up to 40% of requests to list or create agent sessions failed. Copilot Chat on GitHub.com showed degraded agent session features during this time. The issue occurred when a large increase in task-list requests from a client experiment overloaded the cache that supports the agent sessions service. We mitigated the impact by increasing cache capacity, reducing nonessential load on the cache, and disabling the client experiment. We are moving rate limiting to separate infrastructure, reducing client polling, and improving our alerting so we can detect and respond to this type of overload faster.