Incident with Copilot
This summary is created by Generative AI and may differ from the actual content.
Overview
On 2026-08-03 between 06:52 and 11:25 UTC, GitHub Copilot users experienced errors when using chat and agent features. Requests to list available models failed, causing chat and agent interactions to fail since every interaction begins by retrieving the model list. Approximately 3% of model-listing requests failed on average, with failures approaching 100% during peak traffic periods. Approximately 4,066 users were affected in a single 60-minute window, concentrated among IDE-based clients. The underlying AI models remained healthy throughout the incident.
Impact
Approximately 4,066 users affected in a single 60-minute window, concentrated among IDE-based clients. On average about 3% of model-listing requests failed, with failures significantly higher during peak-traffic periods, at times approaching 100% for affected internal lookups.
Trigger
An increase in how often clients requested the model list, which pushed an internal user-authorization lookup past a rate limit. The rate-limited responses were surfaced to users as errors.
Detection
User reports of degraded performance triggered investigation. The incident was initially detected through reports rather than proactive monitoring, with investigation beginning at 09:53 UTC.
Resolution
Mitigated impact by increasing how long Copilot caches the authorization lookup, which reduced load on the internal service. Additional capacity and rate-limit changes were put in progress. Monitoring was maintained to ensure stability.
Root Cause
An increase in client request frequency for the model list pushed an internal user-authorization lookup service past its rate limit. The rate-limited responses were surfaced to users as errors instead of being handled gracefully.