Docker outage history

During a planned database maintenance, four Docker Hub components were placed into maintenance mode: Docker Hub Registry, Docker Hub Web Services, Docker Package Repositories, and Docker Hub Automated Builds. The incident began at 11:15 AM UTC as the database backup for the upcoming update took longer than anticipated, causing degraded access while the backup completed.

The status page indicated these components remained in maintenance until the window closed, and services returned to normal by 12:30 PM UTC as the backup finished and updates completed.

During a scheduled database maintenance window, Docker Hub Registry, Web Services, Package Repositories, and Automated Builds were placed into maintenance mode. The window ran from 10:05 AM UTC to 11:05 AM UTC, after which all components returned to normal operation. The Docker status page documented the maintenance start and subsequent return to service, and users may have experienced brief disruptions or degraded performance during image pulls, pushes, builds, and registry interactions.

The incident affected Docker Hub Registry and Docker Hub Web Services, causing issues with viewing newly pushed images and tags. The issue began at 1:30 PM UTC and was resolved by 1:40 PM UTC, lasting about ten minutes.

Build workflows for DHI customizations in Docker were affected as upstream outages blocked the process. The degradation began at 3:25 PM UTC when Docker Hardened Images reported issues tied to GitHub's outage, and continued due to a separate third-party vendor outage. The degraded state persisted until 10:10 PM UTC when the status returned to normal. This window likely caused interruptions or slower processing for workflows depending on these builds.

Users relying on Docker Hub Automated Builds to trigger on Bitbucket pushes experienced a partial outage affecting this workflow. Bitbucket-based builds did not start, while GitHub-powered builds continued to run, likely delaying CI pipelines and image updates for Bitbucket repositories. The impact was limited to Bitbucket-triggered builds and did not affect other trigger sources.

Timeline: Impact began at 2:15 PM UTC with a degraded signal. Engineers identified the issue by 2:35 PM UTC and deployed a fix by 3:45 PM UTC. The service returned to normal for Bitbucket builds by 4:15 PM UTC after monitoring.

Docker Hub experienced a tag visibility issue for a subset of tags pushed on July 29. Those tags were pullable but not visible in the Hub UI for pushes around 4:20 PM UTC to 5:30 PM UTC, impacting Docker Hardened Images, Docker Hub Registry, and Docker Hub Web Services. The problem persisted across July 30 and into July 31, with multiple status updates describing backfilling of missing entries and UI inconsistencies during the restoration process. By 4:00 PM UTC the three core components returned to normal and backfill of tag listings continued until the UI inconsistency was resolved.

Users relying on Docker Hub email notifications were unable to receive emails when a third-party email provider experienced issues. Docker Hub Web Services reported degraded email delivery impacting outgoing messages, while the team worked on workarounds. The disruption began at 7:20 PM UTC and was resolved by 7:50 PM UTC as the vendor recovered, after which email delivery returned to normal.

From 8:35 AM UTC to 9:30 AM UTC, the DHI Catalog failed to display CVE details on the vulnerabilities tab for Docker Hardened Images, causing degraded visibility of security advisories for users reviewing vulnerabilities. This partial outage affected CVE visibility but did not indicate a broader service disruption. The incident lasted about 55 minutes and the rest of Docker's offerings remained available.

The Docker team issued four official updates during the incident. At 8:45 AM UTC a fix was being rolled out, at 8:55 AM UTC the fix was deployed and monitoring began, and by 9:30 AM UTC the CVE display had been restored and the Docker Hardened Images component returned to normal.

Users attempting to sign in to Docker Desktop faced authentication failures and sign-in delays as the Docker Authentication component encountered problems. The issue began at 6:10 PM UTC and was resolved by 6:50 PM UTC, lasting about 40 minutes.

The Docker status page documented an active investigation and later indicated the component had returned to normal operation.

An outage disrupted Docker Desktop downloads from desktop.docker.com and updates for about 1 hour 20 minutes, preventing users from installing or updating Docker Desktop. The issue was linked to a dependent service and briefly degraded the Docker Package Repositories component, affecting user attempts to obtain or apply Docker components.

The incident began at 9:40 PM UTC and was resolved by 11:00 PM UTC, lasting 1 hour 20 minutes. The Docker status page posted two updates during the incident, describing the dependency problem and the mitigation effort as the team worked toward restoration.