New B200 spot capacity is live in US East from $1.69 per GPU-hour. See availability
Incident history

August 2025.

Every incident and maintenance window that started this month, newest first, with what we posted at the time. Times are UTC. The record starts at launch and nothing is removed from it.

Incidents
2111 minutes of degraded or reduced service
Major outages
0None
Maintenance windows
1Announced at least seven days ahead
Record since
14 Sep 2024Platform launch; every event since is listed

Metadata endpoint slow to respond

Resolved
Degraded performance22 Aug 2025, 08:04 UTC · lasted 23 minInstance metadata service

Resolved — This incident has been resolved.

API returning errors for all requests

Resolved
Partial outage14 Aug 2025, 12:17 UTC · lasted 1 h 28 minAPI

Resolved — This incident has been resolved. A post-incident review will be published here.

Storage cluster firmware update — EU Central

Completed
Scheduled maintenance6 Aug 2025, 01:00 UTC · lasted 3 h 12 minPersistent storage — EU Central (Frankfurt)

Completed — The maintenance is complete and every component is back to normal operation.

Uptime counts major outages in full and partial outages at 60% of their duration; degraded performance and scheduled maintenance are not counted as downtime. Feeds: Atom · JSON.