Home / BeaverDeck / Docs / Insights Guide / GPU Insights / GPU Node Scheduling
GPU Node Scheduling
BeaverDeck uses this check to identify a specific gpu condition that may need operator review.
| Check type | gpu-node-cordoned |
|---|---|
| Insights section | GPU Insights |
| Alert severity | Warning |
When It Reports A Finding
A Node advertises GPU capacity and is marked unschedulable.
Why This Is A Problem
The cluster owns expensive GPU capacity that cannot accept new pods, reducing available headroom and potentially blocking GPU workloads.
Recommended Response
- Confirm whether the node is intentionally cordoned for maintenance or incident response.
- Complete the maintenance or drain procedure and verify GPU driver and node health.
- Uncordon the node only when it is safe to receive workloads.
Scope And Limitations
A cordoned GPU node during planned maintenance is expected. The check does not determine whether the maintenance state is justified.
After remediation: refresh GPU Insights and verify the underlying
resource or metric. Suppress the finding only when the condition is intentional and its risk is accepted.