GitHub Outage History

GitHub users experienced degraded performance for Git Operations for a brief period, impacting push, fetch, and clone workflows. The disruption began around 2:00 PM UTC, with a noticeable uptick in reports by 2:15 PM UTC. GitHub acknowledged the issue at 2:26 PM UTC with an update noting high errors on Git Operations and that engineers were investigating degraded performance. The problem was resolved by 2:36 PM UTC as Git Operations returned to normal. A sample user report from Raleigh, NC appeared around 2:35 PM UTC, illustrating the real-time impact observed by developers. The overall impact remained limited to Git operations without broader service disruption.

GitHub users faced an outage that disrupted pull request workflows after CCR and CCA failed to start for PR comments, causing degraded performance for some GitHub services. The incident began at 5:06 AM UTC and persisted for nearly two hours as engineers investigated.

Developers were unable to create new Pull Request threads, including line and file comments. Responses to existing PR threads remained functional, and top-level PR comments, opening/merging PRs, and Actions continued to work. GitHub noted the issue and indicated that they identified the cause and began mitigation after the problem was detected.

Impact began at 3:26 PM UTC and degraded through the afternoon, with a partial outage noted around 4:11 PM UTC. A mitigation was applied and the team targeted full recovery by 8:00 PM UTC. Partial recovery appeared by 5:56 PM UTC and full recovery followed by 7:06 PM UTC. User reports suggest widespread activity and issues with PR thread creation across multiple regions, while other PR functionality remained usable.

GitHub users reported a brief outage affecting Copilot Cloud Agent sessions and the ability to view them, with some GitHub services briefly impacted. The disruption began at 11:26 AM UTC and was resolved by 12:01 PM UTC, lasting about 35 minutes. A small number of user reports appeared between 11:30 AM UTC and 11:45 AM UTC from different regions, describing failures when starting Copilot Cloud Agent sessions or loading session details.

GitHub's status page acknowledged the issue with an update noting a disruption with some GitHub services and specifically that investigators were looking into the ability to start Copilot Cloud Agent sessions and view them. After the incident was marked resolved, services returned to normal by 12:01 PM UTC. There was no Early Warning Signal for this incident.

Users relying on GitHub Actions workflows faced degraded availability beginning at 7:21 AM UTC as GitHub flagged an incident affecting Actions and noted issues with ubuntu standard hosted runners leading to long queue times. The team continued investigating and working on a fix as queues and execution delays affected many workflows. By 9:11 AM UTC, a mitigation had been applied to address the queuing and running issues, and telemetry indicated improvements while engineers monitored for full recovery.

The degradation affecting Actions was mitigated by 9:21 AM UTC, with the Actions component returning toward normal operation, and full recovery followed shortly after. By 9:46 AM UTC the service was reported as up again. Across many repositories, developers reported long wait times and difficulty running Actions workflows, but the service was restored as the incident concluded.

From 1:41 PM UTC, GitHub Actions began experiencing degraded performance as queue times increased for Jobs on Standard Hosted Runners in East US, impacting a portion of runs and causing delays and some failures. Hosted Runners with Private Networking could fail over to a different Azure region to mitigate the impact. The issue progressed into the afternoon as engineers worked with the compute provider to restore capacity.

Separately, SSH-based Git operations showed elevated latency and failures between approximately 2:00 PM UTC and 4:10 PM UTC, while HTTP-based Git operations were largely unaffected. Mitigations were deployed through the afternoon, with signs of recovery for Standard Hosted Runners by 4:36 PM UTC and ongoing work to prevent further impact. By 6:26 PM UTC the degradation affecting Git Operations had been mitigated and monitoring continued; the status returned to up around 6:36 PM UTC.

A GitHub incident began around 3:46 PM UTC, with degraded performance first reported for Issues and Webhooks. Over the next while, additional areas across GitHub services were affected, including Git Operations, Actions, Packages, Pull Requests, and Pages. Some users experienced problems submitting PR reviews, such as errors mentioning invalid JSON, as the issue evolved.

The status updates describe a cascade of degraded components, with services moving from normal to degraded and then recovering. Recovery began for Git Operations around 4:26 PM UTC and for Pages around 4:31 PM UTC, followed by Webhooks and Issues returning to normal by 4:41 PM UTC. Overall, the incident flowed through multiple parts of the platform before restoring typical functionality.

The impact appeared global, with user reports from regions including the United States, Pakistan, and Kazakhstan, indicating broad disruption to workflows reliant on issues, PRs, webhooks, and automation. The window spanned roughly an hour from the initial degradation to full restoration, with users across regions experiencing degraded performance and intermittent errors during the event.

GitHub users faced a broad disruption affecting Actions hosted runners and the visibility of pull requests. Hosted Actions runners for ubuntu-latest and ubuntu-24.04 ran into capacity constraints, yielding longer queue times and occasional failures to start jobs. Simultaneously, search and PR pages relying on Elasticsearch displayed degraded performance or incomplete results, reducing visibility into open and merged PRs. Impact began at 2:01 PM UTC and was felt across many repositories globally.

Status updates described mitigations and progressive recovery. A mitigation to unblock running Actions was applied at 3:21 PM UTC, while Elasticsearch indexing work progressed with indexes being reindexed around 4:01 PM UTC. By 3:46 PM UTC, less than 2% of hosted ubuntu-latest and ubuntu-24.04 runs were delayed, and Actions began restoring on affected runners by 5:11 PM UTC. Throughout 28–29 April, the team continued reindexing and mitigation efforts with guidance that full PR listings would recover within roughly 24 hours.

Recovery progressed over the next days: by 10:26 PM UTC, search/indexing functionality was restored for over 99% of impacted pull requests; by 3:51 AM UTC missing search records were repaired. The Pull Requests component moved to up on 4:16 AM UTC, signaling near-full restoration of PR visibility as the incident concluded.

GitHub experienced a broad degradation beginning Apr 27, 2026 4:31 PM UTC as search capabilities were slower and investigation began. The issue extended beyond search to multiple GitHub services, with workflow runs failing, projects failing to load, and PRs not loading or displaying. By 6:21 PM UTC, Pull Requests showed degraded availability and the impact spread to other areas including Packages, Issues, and Codex agent runs.

Engineers later identified the cause as an elevated load on ElasticSearch clusters, which stressed search and cascaded into other services. The team disabled the offending load and moved to remediation. By 7:51 PM UTC, the source of the additional load had been identified and actions were underway to restore stability. By 9:36 PM UTC, the affected components were recovering, with PRs, Actions, Packages and Issues returning toward availability, though the overall service remained degraded. The incident was fully mitigated and service restored by 10:51 PM UTC, with ongoing monitoring to ensure stability. Reports indicated the disruption appeared to have global reach across multiple regions.

Impact began at Apr 24, 2026 7:06 PM UTC when GitHub flagged delays with Actions Jobs for Larger Runners using VNet Injection in the East US region. The issue led to degraded performance for workflows and related services, with updates noting ongoing investigation and a broader public impact tied to Azure East US workloads. Users across regions reported problems ranging from difficulty connecting to Copilot to language model availability issues, indicating the incident affected developer tooling and CI workflows in the affected region.