Status
All systems operational.
Checked every 60 seconds from four regions. We publish the raw numbers rather than a reassuring green badge.
| Component | State | 30-day uptime | p95 latency |
|---|---|---|---|
| Upload endpoint | Operational | 99.98% | 142 ms |
| Download edge | Operational | 99.99% | 38 ms |
| REST API | Operational | 99.97% | 91 ms |
| Web dashboard | Operational | 99.95% | 210 ms |
| Billing | Degraded | 99.71% | 480 ms |
Billing is slow while we migrate invoice generation to a new worker pool. Uploads and downloads are unaffected.
Recent incidents
29 September 2026 — elevated upload errors (47 minutes). A storage node in the Helsinki region ran out of inodes after a batch of very small chunks. Roughly 2% of uploads returned 500 until the node was drained. Chunked uploads resumed automatically; single-shot uploads had to be retried by hand. We now alert on inode headroom, not just disk bytes.
11 September 2026 — download links slow in South America (3 hours). A transit provider withdrew a route and traffic took a longer path through Miami. No errors, but p95 download latency tripled. We added a second transit provider for that edge.
2 August 2026 — scheduled maintenance (22 minutes). Database major version upgrade. Uploads were paused and queued; nothing was lost.
Subscribe
Incident notices go out by email and on this page. Write to
hello@ditrib.online to be added to the list, or poll
GET /v1/health if you would rather automate it.