Buffer

Help CenterSubscribe to updates
Powered by
Privacy policy

·

Terms of service
Write-up
Buffer Stuck Loading
Partial outage
View the incident
Summary

On 28 May 2026, Buffer's Publish web app became unavailable due to a critical startup error introduced by a recent code change. The issue affected all users attempting to access the application, with approximately 2,000 user accounts impacted during the outage window. The issue has been fully resolved and the application is operating normally.

What Happened

At 17:35 BST, we identified that the Publish web app was stuck in a loading loop and unavailable to all users. Our engineering team immediately began investigating and identified that a recent dependency update had introduced a Rolldown runtime error during application bootstrap.

Impact

During the outage:

  • All users were unable to access the Publish web app

  • Approximately 2,000 user accounts attempted to load the application and experienced an infinite loading spinner

  • The Content Creation and Advocacy teams were unable to access publishing functionality

Resolution

Our engineering team quickly identified the faulty code change responsible for the issue and reverted it. A new deployment was initiated, and by 17:49 BST, the Publish web app was loading correctly for all users. We confirmed the fix was stable and monitored for any additional issues. No further problems were reported after the deployment.

Root Cause

The incident was caused by an architecture issue introduced during technical clean-up work. A dependency update included in a recent pull request inadvertently introduced a Rolldown runtime error that prevented the application from starting correctly.

What We're Doing Next

We are implementing several improvements to prevent similar incidents:

  • Pinning all dependencies across the project to prevent unexpected updates

  • Adding lockfile checks to our vulnerability investigation processes

  • Enhancing CI visibility to show which services are affected by code changes

  • Implementing additional monitoring and synthetic checks to catch startup failures faster

  • Reviewing our deployment notification systems to improve incident response times