Skip to content
GitHub · Developer toolsAug 20, 2026, 14:43 UTC

Intermittent failures creating agent tasks

CriticalDeploymentUpdated 5h ago
Aug 20, 14:43 UTCAug 21, 00:37 UTC
Duration
9h 54m
Impact
Critical
Root cause
Deployment
GitHub, 90 days
68 incidents
Affected
Not listed by the vendor.

Lesson: Teams should track agent task creation pipelines to handle intermittent failures promptly.

What happened

Between 13:57 UTC on August 20 and 00:37 UTC on August 21, 2026, some users of the Copilot Cloud Agent experienced delays of up to 60 to 90 minutes in seeing the status and results of their agent tasks. The agent tasks themselves continued to run and complete during this time; only the visibility of their status was delayed. The cause was a regional outage in a third-party cloud database service that Copilot uses to store agent task status. We failed over the affected database to a healthy region, added processing capacity to work through the backlog, and restored normal operation once the underlying service recovered. No task data was lost during the incident. To prevent repetition of similar incidents, we are removing the database configuration that made us vulnerable to this regional outage and improving our database failover procedures.

Timeline

  1. Resolved · Aug 21, 00:37 UTC
    Between 13:57 UTC on August 20 and 00:37 UTC on August 21, 2026, some users of the Copilot Cloud Agent experienced delays of up to 60 to 90 minutes in seeing the status and results of their agent tasks. The agent tasks themselves continued to run and complete during this time; only the visibility of their status was delayed. The cause was a regional outage in a third-party cloud database service that Copilot uses to store agent task status. We failed over the affected database to a healthy region, added processing capacity to work through the backlog, and restored normal operation once the underlying service recovered. No task data was lost during the incident. To prevent repetition of similar incidents, we are removing the database configuration that made us vulnerable to this regional outage and improving our database failover procedures.
  2. Investigating · Aug 20, 20:37 UTC
    We are seeing gradual recovery in Copilot Cloud Agent task status visibility as we deploy a fix for the root cause. Session output remains delayed by approximately one hour while remediation continues.
  3. Investigating · Aug 20, 19:35 UTC
    We are continuing to observe gradual recovery for Copilot Cloud Agent task status visibility. Session output continues to be delayed by approximately 1 hour as our remediation steps take effect.
  4. Investigating · Aug 20, 18:45 UTC
    We are continuing to observe gradual recovery for Copilot Cloud Agent task status visibility, with session output delayed by approximately 1 hour. We have taken additional steps to accelerate the recovery and expect this to take effect within the next hour.
  5. Investigating · Aug 20, 18:04 UTC
    We are continuing to observe gradual recovery for Copilot Cloud Agent task status visibility, with session output delayed by approximately 1 hour. We have taken additional steps to accelerate the recovery and are continuing to monitor the impact.
  6. Investigating · Aug 20, 17:32 UTC
    We are observing gradual recovery for Copilot Cloud Agent task status visibility, with session output delayed approximately 1 hour. We have taken additional steps to accelerate the recovery and are continuing to monitor the impact.
  7. Investigating · Aug 20, 17:05 UTC
    We are seeing signs of recovery for Copilot Cloud Agent task status visibility, but this recovery is slower than anticipated. We are pursuing additional mitigating measures to accelerate recovery.
  8. Investigating · Aug 20, 16:14 UTC
    Users are experiencing delays when starting tasks using Copilot Cloud Agent and are not be able to see the status of these tasks. Copilot Cloud Agent tasks are still being completed. We have identified the cause of the issue and are putting mitigations in place to return service to normal levels. We will provide another update about the expected recovery time shortly.
  9. Investigating · Aug 20, 15:41 UTC
    We are experiencing issues with Copilot Cloud Agent tasks, resulting in newly started tasks not properly displaying on-going progress. These Copilot Cloud Agent tasks are still being completed correctly but lack proper visibility. We are actively investigating the issue and will provide updates as we learn more.
  10. Investigating · Aug 20, 15:01 UTC
    We have identified the problematic component and are working to fail over to a healthy instance. Further updates will be provided as we perform mitigations.
  11. Investigating · Aug 20, 14:51 UTC
    Users may experience delays when starting tasks using Copilot Cloud Agent. We are actively investigating the issue and will provide updates as we learn more.
  12. Investigating · Aug 20, 14:43 UTC
    We are investigating reports of impacted performance for some GitHub services.

More from GitHub

Full history
StartedIncidentDuration
Sep 2310:11 UTCIncident across several servicesOngoing
Sep 2022:13 UTCIncident with Pull Requests1h 9m
Sep 1720:59 UTCElevated rate of errors for OpenAI models provided by Copilot50m
Sep 1607:20 UTCDegradation with Gemini 3.8 Flash10h 28m
Sep 1519:11 UTCDisruption with some GitHub services49m
Sep 1509:47 UTCDisruption with some GitHub services1h 30m

Also caused by deployment or rollout

All
StartedIncidentDuration
Sep 1106:22 UTCIssues with Workers VPC hostname route resolution on 2026-09-11Cloudflare0m
Aug 2000:31 UTC[Medium] Issue with Microsoft Office Integration and Box EditBox1h 13m
Aug 622:22 UTCTrouble Using Search Bar For Some AdminsSlack2h 44m
Jul 2110:23 UTCJob runs failing at the git clone stepdbt Labs2h 29m
Jul 1907:33 UTCSome EMEA customers experiencing Git clone failures (403 errors) on thier accountsdbt Labs6h 22m
Jul 200:17 UTCTraces, spans, logs inaccesible for query or ingestion in de and usSentry3h 39m

From vendors' own status pages and disclosures. Times as reported. Logos via logo.dev; trademarks belong to their owners.

Weekly: the week's major outages, postmortems and breaches, Saturday mornings.