Status: Resolved

This incident was caused by a platform-wide outage on Meta's side, not by any issue within Buffer. Meta experienced a significant disruption affecting their APIs, webhook delivery systems, and consumer-facing web apps — Facebook, Instagram, and Threads — globally. The outage was severe enough that even Meta's own web products failed to load in browsers.

The Meta-side cause was confirmed by multiple independent sources: several other social media management platforms experienced the same disruption simultaneously, third-party network monitoring services reported elevated error rates for Meta, and Meta later acknowledged the issue on their own status page (https://metastatus.com/graph-api/history).

Because this originated entirely upstream at Meta, it was not preventable through any action on our side.

For approximately 4 hours, 20 minutes, customers were unable to perform any Meta-related actions through Buffer. Three core workflows were affected:

• Publishing — Posts could not be scheduled or published to Facebook, Instagram, or Threads.

• Engagement — Comment ingestion from Meta platforms was disrupted, so customers couldn't view or respond to audience interactions through Buffer.

• Analytics — Data retrieval from Meta platforms was unavailable while their APIs were down.

Because the root cause sat entirely on Meta's side, recovery ultimately depended on Meta restoring their services. In the meantime, our team:

• Posted the disruption to our public status page and surfaced an in-app banner.

• Notified affected customers directly.

• Continuously monitored Meta's recovery across all three platforms.

As recovery progressed, Facebook returned to a >95% success rate first, while Instagram and Threads initially lagged at roughly 25% before climbing. Once all Meta channels were sustained above 98%, we marked the incident resolved and updated the status page accordingly.

Preventability — This was a third-party outage affecting Meta globally, hitting not just Buffer but peer products across the industry at the same time. There was no internal action that could have prevented it.

Early detection worked well — Our automated monitoring surfaced the disruption quickly: throughput from Meta's webhook events dropped sharply and triggered an alert, and our Meta broadcasting alerts fired shortly after. Detection was fast and the signal was clear, which let us communicate with customers early.