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

Instance launches failing in US West (Oregon)

26 launches failed in US West (Oregon) over 22 minutes; running instances were not affected.

Reference cabb7ce644b4 · US West (Oregon) · Updated just now

Status
ResolvedResolved — last update 25 Mar 2026, 08:24 UTC
Started
25 Mar, 08:02Wednesday 25 March 2026 UTC
Duration
22 minEnded 25 Mar 2026, 08:24 UTC
Components
1Instance launch and lifecycle — US West (Oregon)
Timeline

What we posted, as we posted it.

Resolved4 updates · UTC
  1. ResolvedThis incident has been resolved. Failed launches were not billed.25 Mar, 08:24
  2. MonitoringLaunches have been succeeding in US West (Oregon) since 08:18 UTC. The cluster is being brought back with a fresh agent. We are monitoring.25 Mar, 08:18
  3. IdentifiedOne hypervisor cluster in US West (Oregon) lost its management agent after a storage timeout, so launches scheduled on it failed. The cluster has been taken out of scheduling; 26 launches were affected.25 Mar, 08:06
  4. InvestigatingNew instance launches in US West (Oregon) are failing. Running instances are not affected. Launches in other regions are working.25 Mar, 08:02

Affected components

Details

Type
Partial outage
Started
25 Mar 2026, 08:02 UTC
Ended
25 Mar 2026, 08:24 UTC
Region
US West (Oregon)
Impact
26 launches failed in US West (Oregon) over 22 minutes; running instances were not affected.
Reference
cabb7ce644b4

Follow the Atom feed