Google CloudSenior
GKE OutOfPods: node has reached its maximum pod count
OutOfPods means a node is at its pod capacity. Raise the per-node limit or enable the cluster autoscaler, and mind the startup race with static pods.
What you see
Node didn't have enough resource: pods, requested: 1, used: 32, capacity: 32- Pod status
OutOfPods
What it means
Each node has a maximum number of pods it can run, separate from CPU and memory. When it is full, the scheduler cannot add another pod, even if CPU and memory are free.
Google notes it commonly happens at node startup: the scheduler assigns pods to a new node before the kubelet reports static pods such as kube-proxy, which need their own pod capacity.
How to confirm the cause
- 1Read the event on the failing pod
kubectl describe pod POD_NAME
Look for the
requested,usedandcapacityfigures for pods. - 2Count pods per node
kubectl get pods -A -o wide --field-selector spec.nodeName=NODE_NAME
Compare with the node's pod capacity.
Causes and fixes
The per-node pod limit is too low
- HOW TO TELL
- Nodes consistently reach the limit.
- ROOT CAUSE
- The node pool is set to a maximum pods per node that is small for your pod density.
- FIX
- Increase
--max-pods-per-nodefor the node pool. Larger pod counts may need larger nodes to cover their resource demands.
Not enough nodes for the number of pods
- HOW TO TELL
- Many small pods and few nodes.
- ROOT CAUSE
- Total pod capacity across nodes is below demand.
- FIX
- Enable the cluster autoscaler and node auto-provisioning so nodes are added when capacity runs out.
How to stop it coming back
- Plan pod density when you size node pools.
- Enable autoscaling for node pools.
- Alert on node pod count close to capacity.
Keep going
- Runbook: GKE PodUnschedulable: insufficient resources, MatchNodeSelector, taints
- Runbook: Pod stuck in Pending with FailedScheduling
- Cheat sheet: kubectl
- Cheat sheet: gcloud
- Lesson: the kubernetes topic in Learn
Sources
Google Cloud documentation (docs.cloud.google.com), read 18 Nov 2026.