On the evening of May 18, Bluesky had a platform-wide outage — their API and their own app were both down. Because we publish to Bluesky through their API, posts going out to Bluesky during that window failed. This was entirely on Bluesky's side; nothing on Buffer's end changed.
For roughly three to four hours that evening (US time), posts scheduled to Bluesky failed to publish. Around 433 posts were affected. Everything else in Buffer — other channels, the dashboard, scheduling — kept working normally, and we didn't see any reports come in over email or social.
Steps to resolution
We caught the spike quickly and confirmed it was Bluesky rather than us — their status page was reporting the outage and their app wouldn't load for any of us either.
We put a notice on our status page so people knew what was happening, then kept an eye on things until Bluesky restored service.
Once they recovered, posting went back to normal within a few minutes and we let it run stable for half an hour before calling it resolved.
Key learnings
There wasn't much we could change here — the fix was on Bluesky's side. The response went smoothly, with a clean handoff across timezones overnight.
The one thing we're chewing on is that when a platform like Bluesky goes down, every failed post surfaces immediately rather than being retried, so we're looking at whether smarter retry handling would soften short outages like this one in future.
