BeaverDeck logo
BeaverDeck

Home / BeaverDeck / Docs / Insights Guide / GPU Insights / GPU Pod Requests

GPU Pod Requests

BeaverDeck uses this check to identify a specific gpu condition that may need operator review.

Permissions: viewing checks requires insights: view. Opening a linked object or logs requires the corresponding resource permission, and the BeaverDeck ServiceAccount must be allowed to read the Kubernetes resources used by the check. Suppressing a finding requires insights: edit and affects all users.
Check typegpu-pod-requests
Insights sectionGPU Insights
Alert severityWarning

When It Reports A Finding

An active GPU-requesting Pod has an init or application container without a CPU request or memory request.

Why This Is A Problem

The GPU can be reserved while supporting CPU or memory is scheduled inaccurately, leading to poor placement, node contention, and wasted accelerator time.

Recommended Response

  1. Measure CPU and memory needed to feed the GPU workload and initialize models or data.
  2. Set requests for every listed container in the owning workload.
  3. Review GPU, CPU, memory, and replica scaling together so one resource does not strand the others.

Scope And Limitations

The check verifies request presence, not whether values are right-sized. It uses the same CPU and memory request rules as Workload Insights.

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.