← Back to the carnage
FLY

⚪ 🔐 Vault cluster went catatonic, turning certificate operations into an infinite timeout buffet. API requests decided to just hang out indefinitely while TLS connections waited for Godot. All better now, but that was 12 minutes of pure infrastructure theater. (Auto-resolved: no updates in 24h)

⏰ Mar 03, 2026 Fixed (for now) spotted via status_rss

// The official story

Certificates issues affecting API and proxy

// Status page says

Mar 3 , 00:54 UTC Resolved - Between 19:54 and 20:06 UTC, our Vault cluster serving app certificates was unavailable. This caused various API requests to fail, mainly operations on certificates but also app creates and IP assignments. As the failure mode was Vault requests hanging rather than failing immediately, TLS requests through fly-proxy for domains where the certificate was not cached on the local node remained open

🔗 Official status page (for the PR-approved version)

https://status.flyio.net/incidents/f32jccgkcl0p

Fly's Greatest Hits

🟢 FLY 1d ago
⚪ 🎭 Turns out SSL certificates have commitment issues too. Fly's sprite subdomain decided to play hard-to-get with a certificate error, leaving developers staring at broken image tags for an hour. Romance was rekindled at 00:45 UTC. (Auto-resolved: no updates in 24h)
🟢 FLY 4d ago
🟠 Fly's sprite creation engine decided Europe was the only continent worth serving. Everyone else gets to experience the joy of failed deployments across all regions. Investigating why the infrastructure gods chose geographic discrimination today. (Auto-resolved: no updates in 24h)
🟢 FLY 5d ago
🔴 Fly's LHR datacenter discovered BGP is hard. Regional anycast and static IPs went dark while traffic gets rerouted to the continent like some kind of digital Brexit. Engineers identified the problem, which is tech-speak for 'we found where the fire is.'. (Auto-resolved: no updates in 24h)
🟢 FLY Oct 03, 2026
🟠 Fly's Amsterdam region decided to take an unscheduled nap. Management plane and AMS hosts are stumbling back online while engineers frantically assess the damage. MPG impact assessment in progress—translation: they're still counting the bodies. (Auto-resolved: no updates in 24h)