Scheduled Maintenance: Global router upgrade and Nebius image rollouts
resolved- Started
- Aug 3, 3:00 AM EDT
- Resolved
- Aug 3, 7:00 AM EDT
- Duration
- 4h
Cerebrium's updates (3)
- update · Aug 3, 3:00 AM EDT
We are upgrading our global router to improve latency and routing decisions, and rolling out new machine images across our Nebius regions to enable checkpointing. Requests routed through the global router may see increased latency or transient errors while the upgrade rolls out. Nodes in Nebius eu-north1 and us-central1 will be rolled onto the new image. Active runs in these regions may be interrupted. Regions outside Nebius are not affected by the image rollout. Deployments and inference requests may still see elevated latency across all regions while the router work completes. If you run latency-sensitive production traffic, plan for retries during the window. We will post updates here as the work progresses.
- update · Aug 3, 5:58 AM EDT
Maintenance is taking longer than anticipated and will continue for an additional hour. As team rolls out updates we do our best to ensure that services are not disrupted. Thank you for your patience and understanding
- update · Aug 3, 7:00 AM EDT
Maintenance ended
What our checks recorded
We checked Cerebrium 48 times from Cerebrium's start time to its resolution, once every five minutes. Every one of them reached Cerebrium's site, so this incident did not show up as a failed page load. Our check loads the public site; it cannot see an API error or a broken feature. Median response time was 186 ms, against 165 ms over the 24 hours before. Cerebrium posted the incident at Aug 3, 3:00 AM EDT.
Updates are Cerebrium's own words from its status page (status: resolved); we never add an inferred cause.