Resolved incident

Pending Tasks on H100

Beam · None · Started

Duration
< 1m
Severity
None
Status
Resolved

Summary

None · Resolved

Earlier today, some of our H100 nodes ran out of disk space which caused them to go offline. If your workloads were scheduled on these nodes, some of your tasks may have timed out during this window. We have taken steps to prevent this from happening again, and we have added credits to the accounts that were affected.

View official incident ↗

Event timeline

How this incident unfolded, from the official status page.

  1. Resolved

    Beam

    Earlier today, some of our H100 nodes ran out of disk space which caused them to go offline. If your workloads were scheduled on these nodes, some of your tasks may have timed out during this window. We have taken steps to prevent this from happening again, and we have added credits to the accounts that were affected.