System status
Availability of every service, measured from outside our network every 30 seconds from eleven probe locations.
Last 90 days
green no failed checks amber degraded performance orange outage. Uptime is the share of probe checks that succeeded and returned in under 2 seconds.
Serving regions
| Region | Location | Status | Uplink | Median tile time | Load now |
|---|---|---|---|---|---|
| eu-north | Tallinn, Estonia | operational | 200 Gbit/s | 29 ms | 31% |
| eu-central | Frankfurt, Germany | operational | 400 Gbit/s | 36 ms | 44% |
| us-east | Ashburn, Virginia | operational | 300 Gbit/s | 52 ms | 38% |
| ap-southeast | Singapore | operational | 160 Gbit/s | 71 ms | 52% |
| sa-east | São Paulo, Brazil | operational | 80 Gbit/s | 64 ms | 27% |
October 2026: 2.1 PB delivered, peak 214 Gbit/s on 21 October at 07:20 UTC, during the Atlas planet refresh window.
Incident history
9 October 2026: elevated errors in Singapore
From 14:02 to 14:16 UTC, 4.2 percent of requests served from ap-southeast returned 502. A router software update on one of two upstream links dropped traffic to a congested path. We drained the region at 14:09 and restored the link at 14:16. Customers in Asia saw higher latency while traffic went to Frankfurt.
15 September 2026: Frankfurt tile brownout
For 22 minutes from 09:48 UTC, tile responses in eu-central took between 1.8 and 4 seconds. A cache node had been rebuilt with a stale index after a disk replacement. We removed the node, rebuilt the index and added a pre-flight check for restored nodes.
27 August 2026: nightly build four hours late
Extract builds finished at 08:05 UTC instead of 04:15. The OSM replication feed we read from was delayed. Files were correct, but anyone polling before 08:05 saw the previous day. Builds now alert when the input state is more than 30 minutes behind, and the manifest is only written when the date matches.
2 July 2026: certificate renewal failure
One certificate failed to renew on the Ashburn edge, and downloads over HTTPS failed for 6 minutes from 03:12 UTC. Renewal now runs 30 days before expiry and fails loudly.
Subscribe to updates by writing to status@n72.assetlines.net. Incidents lasting more than 5 minutes always get a written report within 3 business days.