Status
System status
All systems operational
API
99.99% uptime
90 days ago
Today
Console
100% uptime
90 days ago
Today
Training clusters · Europe
100% uptime
90 days ago
Today
Training clusters · North America
99.97% uptime
90 days ago
Today
Inference endpoints
99.98% uptime
90 days ago
Today
Object storage
99.96% uptime
90 days ago
Today
Regions
Seven regions, one API
Typical round-trip time from western Europe. Pick the region closest to your data, not your desk.
Amsterdam
ams-1 · Forge, Anvil
~18 ms
Frankfurt
fra-1 · Forge, Spark
~22 ms
Dublin
dub-1 · Spark
~28 ms
Virginia
iad-2 · Forge, Anvil
~88 ms
Oregon
pdx-1 · Forge
~140 ms
Singapore
sin-1 · Spark, edge
~165 ms
Tokyo
nrt-1 · Forge
~230 ms
History
Past incidents
Every incident gets a public write-up within five working days. Each one below also marks its day on the uptime bars.
3 days ago · Inference endpoints
Slow cold starts in sin-1
minor
Endpoints scaling from zero took up to 40 seconds for 26 minutes after a config change. Rolled back.
17 days ago · Training clusters · North America
Fabric switch failure in iad-2
major
One spine switch failed and 96 GPUs lost their fabric link for 38 minutes. Jobs restarted automatically from checkpoints.
41 days ago · API
Elevated API errors
minor
About 2% of API calls returned errors for 11 minutes during a database failover.
64 days ago · Object storage
Slow uploads in fra-1
minor
Large uploads were slower than usual for 50 minutes while a storage node rebuilt.