Date: October 27, 2025 Duration: ~3 hours (14:30 – 19:39 UTC+1) Severity: Major
A code change introduced during a technical refactoring caused social media posts to get stuck in "processing" and not be delivered as expected. The issue primarily affected Instagram, but also impacted LinkedIn, Threads, Facebook, and Pinterest posts with media attachments.
A regression was introduced during a routine technical refactoring. A specific code path involved in retrying posts during the broadcasting process was not updated correctly, causing those posts to get stuck rather than being retried and sent. The error was not caught by our type-checking tooling at the time.
Approximately 19,000 posts were affected and failed to be delivered. The impacted post types were:
All Instagram posts
LinkedIn posts with media
Threads posts with media
Facebook posts with media
Pinterest posts with videos
Once the root cause was identified, we reverted the faulty code change and deployed the fix. We then monitored publishing data over the following hour to confirm that no new posts were getting stuck before declaring the incident resolved.
We identified an opportunity to improve type safety in our internal event-sending mechanism to prevent similar issues from going undetected by our type-checking tools.
We recognized the need for better monitoring on the number of posts stuck in processing, which would allow us to detect and respond to similar issues faster — ideally before customers are impacted.
