Sovereign AI Cloud Pricing How to AI Blog About
System Status

Live platform status

Real-time health for every part of the Sovereign Cloud, 90 days of uptime history, and an honest incident log. When something breaks, it shows up here first.

Operational Degraded performance Outage 90-day history · hover a bar for detail

Get notified about incidents

We'll email you when an incident opens and when it's resolved. EU-resident, unsubscribe anytime.

Incident history

Past 90 days

Every incident, with the timeline and what we changed afterwards.

9 May 2026

Elevated inference latency — Frankfurt Resolved Degraded · 42 min

A subset of model-inference requests in the Frankfurt region saw elevated latency after a GPU node degraded. Requests succeeded but were slower than normal.

15:31 CETResolved — node drained and replaced; latency back to baseline.
15:02 CETIdentified — single GPU node degraded; traffic being rerouted.
14:49 CETInvestigating — monitoring detected raised p95 latency in Frankfurt.
22 Apr 2026

Scheduled DR failover test — Stockholm Maintenance

Planned disaster-recovery exercise validating Stockholm failover. No customer impact; announced five days in advance.

04:40 CETCompleted — failover and failback verified, RPO/RTO within targets.
03:00 CETStarted — controlled failover initiated during the maintenance window.
30 Mar 2026

Office Stack mail delivery delays Resolved Degraded · 26 min

Outbound mail from the Office Stack queued behind a backed-up worker pool. Messages were delivered with delay, none lost.

11:18 CETResolved — queue drained, worker pool scaled and limits raised.
10:52 CETInvestigating — outbound mail queue depth rising in one region.
14 Feb 2026

API rate-limit errors for a subset of accounts Resolved Degraded · 19 min

A configuration change applied stricter rate limits than intended, causing 429 responses for some Agent API callers. Rolled back promptly.

09:27 CETResolved — change rolled back; affected accounts confirmed healthy.
09:08 CETIdentified — recent limit config change linked to the errors.
Before 14 Feb 2026

No incidents reported Operational

All systems operated within targets for the remainder of the 90-day window.