GitHub availability report: March 2026
Source: GitHub Blog
March Incident Summary
In March, we experienced four incidents that resulted in degraded performance across GitHub services.
March 03 18:59 UTC (lasting 1 hour 10 minutes)
Timeframe: 18:46 – 20:09 UTC (March 3, 2026)
Impact: Degraded availability affecting:
github.com- GitHub API
- GitHub Actions
- Git operations (HTTP only; SSH was unaffected)
- GitHub Copilot
- Other dependent services
Key metrics at peak:
| Service | Error / Failure Rate |
|---|---|
github.com requests | ~40 % |
| GitHub API requests | ~43 % |
| Git over HTTP | ~6 % |
| Git over SSH | 0 % |
| GitHub Copilot requests | ~21 % |
| GitHub Actions | < 1 % |
Root cause: A bug introduced while deploying a change to reduce the load on the user‑settings caching mechanism. The bug caused every user’s cache to expire, be recalculated, and rewritten, generating a massive write volume. Replication delays cascaded to all dependent services.
Mitigation: Immediate rollback of the faulty deployment.
Next steps:
- Added a killswitch and improved monitoring for the caching mechanism to detect issues before they affect users.
- Moved the cache mechanism to a dedicated host, isolating future problems to the services that rely on it.
March 05 16:35 UTC (lasting 2 hours 55 minutes)
Timeframe: 16:24 – 19:30 UTC (March 5, 2026)
Impact: GitHub Actions degradation.
- 95 % of workflow runs failed to start within 5 minutes (average delay ≈ 30 minutes).
- 10 % of workflow runs failed with an infrastructure error.
Root cause: Redis infrastructure updates introduced incorrect configuration changes to the Redis load balancer, routing traffic to the wrong host and triggering two incidents.
Mitigation:
- Corrected the misconfigured load balancer (jobs resumed successfully at 17:24 UTC).
- Rolled back the problematic updates and froze further changes in this area pending follow‑up work.
Ongoing work:
- Improve automation to prevent propagation of incorrect configuration changes.
- Enhance alerting to catch misconfigured load balancers earlier.
- Update the Redis client configuration in Actions to better tolerate brief cache interruptions.
March 19 13:44 UTC (lasting 48 minutes)
Timeframes:
- 01:05 – 02:52 UTC (March 19, 2026)
- 00:42 – 01:58 UTC (March 20, 2026)
Impact: Copilot Coding Agent service degradation – users could not start new Copilot Agent sessions or view existing ones.
| Incident | Avg. error rate | Peak error rate |
|---|---|---|
| March 19 | ~53 % | ~93 % |
| March 20 | ~99 % | ~100 % (with significant retry amplification) |
Root cause: An authentication issue prevented the service from connecting to its backing datastore.
Mitigation: Rotated the affected credentials, restoring connectivity. The first incident was fully mitigated by 01:24 UTC; the second occurred because the first remediation was incomplete.
Future safeguards:
- Implemented automated monitoring for credential‑lifecycle events.
- Improved operational processes to reduce detection and mitigation times for similar issues.
March 24 16:59 UTC (lasting 2 hours 52 minutes)
Timeframe: 15:57 – 19:51 UTC (March 24, 2026)
Impact: Microsoft Teams Integration and Teams Copilot Integration services were degraded, preventing delivery of GitHub event notifications to Microsoft Teams.
- Average error rate: 37.4 %
- Peak error rate: 90.1 % (≈ 19 % of all integration installs failed to receive notifications)
Root cause: Outage at an upstream dependency, which produced HTTP 500 errors and connection resets for the Teams integration.
Mitigation: Coordinated with the upstream service teams; the issue was resolved at 19:51 UTC when the upstream incident was mitigated.
Future work: Updating observability and runbooks to shorten mitigation times for similar incidents.
Stay Informed
- Follow our status page for real‑time updates and post‑incident recaps.
- Learn more about our ongoing work in the engineering section of the GitHub Blog.
Written by
Related posts
(Links to related posts would be listed here.)
Explore more from GitHub
Docs
Everything you need to master GitHub, all in one place.
GitHub
Build what’s next on GitHub, the place for anyone…
Start building
Customer stories
Meet the companies and engineering teams that build with GitHub.
The GitHub Podcast
Catch up on the GitHub podcast, a show dedicated to the topics, trends, stories and culture in and around the open‑source developer community on GitHub.