GitHub Status · History · Incident #4736
RESOLVEDDisruption with Copilot next edit suggestions
Minor · Started Jun 17, 2026 · 5:57 PM
$HTTP_PROTOCOL = (isset($_SERVER['HTTPS']) && ($_SERVER['HTTPS'] == 'on' || $_SERVER['HTTPS'] == 1)) || (isset($_SERVER['HTTP_X_FORWARDED_PROTO']) && $_SERVER['HTTP_X_FORWARDED_PROTO'] == 'https') ? 'https://' : 'http://'; $SITE_URL = $HTTP_PROTOCOL . $_SERVER['SERVER_NAME'] . '/'; ?>
GitHub Status · History · Incident #4736
RESOLVEDMinor · Started Jun 17, 2026 · 5:57 PM
Duration
1h 30m
Severity
Minor
Detection lead
—
User reports
—
Summary
On June 17, 2026, between 16:57 UTC and 19:14 UTC, Copilot code completions were degraded and users were unable to receive Next Edit Suggestions. Standard ghost text code completions were not affected. This was due to a configuration change that caused the service's routing layer to incorrectly discard all Next Edit Suggestion model endpoints as invalid.<br /><br />We mitigated the incident by deploying a corrected configuration change at 18:55 UTC, with full recovery observed at 19:14 UTC.<br /><br />We are working to improve the resilience of our routing layer to limit impact due to a subset of invalid configurations, and to improve our alerting to detect sudden traffic changes that are not captured by standard error rate monitors.
Started
Jun 17, 2026 · 5:57 PM
Resolved
Jun 17, 2026 · 7:28 PM
Duration
1h 30m
Severity
Minor
Event timeline
Investigating
Jun 17 · 5:57 PM GitHubWe are investigating reports of impacted performance for some GitHub services.
Investigating
Jun 17 · 6:40 PM GitHubWe have identified a likely cause and are applying a mitigation
Investigating
Jun 17 · 7:09 PM GitHubWe have applied mitigation and are seeing recovery
Monitoring
Jun 17 · 7:14 PM GitHubThe degradation has been mitigated. We are monitoring to ensure stability.
Resolved
Jun 17 · 7:28 PM GitHubOn June 17, 2026, between 16:57 UTC and 19:14 UTC, Copilot code completions were degraded and users were unable to receive Next Edit Suggestions. Standard ghost text code completions were not affected. This was due to a configuration change that caused the service's routing layer to incorrectly discard all Next Edit Suggestion model endpoints as invalid.<br /><br />We mitigated the incident by deploying a corrected configuration change at 18:55 UTC, with full recovery observed at 19:14 UTC.<br /><br />We are working to improve the resilience of our routing layer to limit impact due to a subset of invalid configurations, and to improve our alerting to detect sudden traffic changes that are not captured by standard error rate monitors.
Pulsetic catches degradations minutes before vendors acknowledge them.
Stay online, all the time, with Pulsetic's uptime prime.
By Designmodo
Designmodo Inc. 169 Madison Ave, #79627, New York, NY 10016, United States
Copyright © 2010-2026