All Articles
troubleshootingawsekskubernetes

EKS Fargate Pod Not Scheduling — Causes and Fixes (2026)

Pods stuck in Pending on EKS Fargate? Here are the 8 most common reasons Fargate pods won't schedule and exactly how to fix each one.

DevOpsBoysApr 14, 20264 min read
Share:Tweet

EKS Fargate is serverless Kubernetes — no nodes to manage. But when pods get stuck in Pending with cryptic errors, debugging feels harder than regular EC2 nodes. Here are the real causes and fixes.

How Fargate Scheduling Works

Fargate doesn't use EC2 nodes. Instead, AWS provisions a micro-VM per pod that matches a Fargate Profile. If no profile matches the pod, it never schedules.

Pod created
     ↓
EKS checks Fargate profiles
     ↓
Profile matches? → Fargate VM provisioned → Pod runs
No match? → Pod stays Pending forever

Cause 1: Fargate Profile Doesn't Match Namespace

The most common reason. Fargate profiles are namespace-scoped.

Symptom:

bash
kubectl describe pod my-pod -n my-namespace
# Events: 0/0 nodes are available: Scheduling is disabled on all nodes.

Fix:

bash
# Check what profiles exist
aws eks list-fargate-profiles --cluster-name my-cluster
 
# Check which namespaces a profile covers
aws eks describe-fargate-profile \
  --cluster-name my-cluster \
  --fargate-profile-name my-profile
 
# Output shows:
# "selectors": [{"namespace": "production"}]
# ← if your pod is in "prod", it won't match

Create a profile for the correct namespace:

bash
aws eks create-fargate-profile \
  --cluster-name my-cluster \
  --fargate-profile-name prod-profile \
  --pod-execution-role-arn arn:aws:iam::123456789:role/AmazonEKSFargatePodExecutionRole \
  --selectors namespace=prod

Cause 2: Label Selector Mismatch

Fargate profiles can filter by labels too. If your profile requires a label the pod doesn't have — no scheduling.

Check profile selectors:

bash
aws eks describe-fargate-profile \
  --cluster-name my-cluster \
  --fargate-profile-name my-profile \
  --query 'fargateProfile.selectors'
 
# Output:
# [{"namespace": "production", "labels": {"fargate": "true"}}]

Fix: Add the required label to your pod spec:

yaml
metadata:
  labels:
    fargate: "true"   # ← must match profile selector

Or remove the label requirement from the Fargate profile if it's not needed.

Cause 3: Missing Pod Execution IAM Role

Fargate needs an IAM role to pull images, write logs, and run the pod.

Symptom:

Pod stuck in Pending. No events. Profile exists.

Check:

bash
aws eks describe-fargate-profile \
  --cluster-name my-cluster \
  --fargate-profile-name my-profile \
  --query 'fargateProfile.podExecutionRoleArn'

Create the role if missing:

bash
# Trust policy
cat > fargate-trust.json << 'EOF'
{
  "Version": "2012-10-17",
  "Statement": [{
    "Effect": "Allow",
    "Principal": {"Service": "eks-fargate-pods.amazonaws.com"},
    "Action": "sts:AssumeRole"
  }]
}
EOF
 
aws iam create-role \
  --role-name AmazonEKSFargatePodExecutionRole \
  --assume-role-policy-document file://fargate-trust.json
 
aws iam attach-role-policy \
  --role-name AmazonEKSFargatePodExecutionRole \
  --policy-arn arn:aws:iam::aws:policy/AmazonEKSFargatePodExecutionRolePolicy

Cause 4: Resource Request Exceeds Fargate Limits

Fargate has fixed vCPU/memory tiers. If your request doesn't match a valid combination, the pod won't schedule.

Valid Fargate CPU/Memory combinations:

vCPUMemory options
0.250.5 GB, 1 GB, 2 GB
0.51 GB – 4 GB (1 GB increments)
12 GB – 8 GB (1 GB increments)
24 GB – 16 GB (1 GB increments)
48 GB – 30 GB (1 GB increments)
816 GB – 60 GB (4 GB increments)
1632 GB – 120 GB (8 GB increments)

Check your pod requests:

bash
kubectl get pod my-pod -o jsonpath='{.spec.containers[*].resources}'

Fix: Use valid combinations:

yaml
resources:
  requests:
    cpu: "0.5"
    memory: "1Gi"    # valid: 0.5 vCPU + 1 GB
  limits:
    cpu: "0.5"
    memory: "1Gi"

Cause 5: Fargate Profile Is Being Created/Deleted

Profiles have a status. If it's not ACTIVE, pods won't schedule.

bash
aws eks describe-fargate-profile \
  --cluster-name my-cluster \
  --fargate-profile-name my-profile \
  --query 'fargateProfile.status'
 
# CREATING / DELETING / ACTIVE / CREATE_FAILED / DELETE_FAILED

Wait for ACTIVE or fix a failed profile:

bash
# Delete failed profile and recreate
aws eks delete-fargate-profile \
  --cluster-name my-cluster \
  --fargate-profile-name my-profile

Cause 6: kube-system Pods on Fargate (CoreDNS)

If you added kube-system to a Fargate profile, CoreDNS needs an annotation to run on Fargate.

bash
# Check CoreDNS deployment
kubectl describe deployment coredns -n kube-system | grep eks.amazonaws
 
# If missing, patch it:
kubectl patch deployment coredns \
  -n kube-system \
  --type json \
  -p='[{"op":"remove","path":"/spec/template/metadata/annotations/eks.amazonaws.com~1compute-type"}]'

Cause 7: DaemonSets Don't Work on Fargate

Fargate doesn't support DaemonSets — there are no nodes. Each Fargate pod is its own VM.

bash
kubectl describe pod my-daemonset-pod -n my-namespace
# Warning: DaemonSet pods cannot be scheduled on Fargate

Fix: Use a sidecar pattern instead, or run DaemonSet components as regular pods.

Cause 8: Subnet Has No Available IPs

Fargate assigns an ENI per pod. If your VPC subnet is exhausted, pods can't start.

bash
# Check available IPs in subnets
aws ec2 describe-subnets \
  --filters "Name=vpc-id,Values=vpc-12345" \
  --query 'Subnets[*].{Subnet:SubnetId,Available:AvailableIpAddressCount}'

Fix: Use /24 or larger subnets for Fargate, or add new subnets to the VPC.

Debug Checklist

bash
# 1. Check pod events
kubectl describe pod <pod-name> -n <namespace>
 
# 2. Check profile status
aws eks list-fargate-profiles --cluster-name <cluster>
aws eks describe-fargate-profile --cluster-name <cluster> --fargate-profile-name <profile>
 
# 3. Check profile matches pod namespace + labels
kubectl get pod <pod-name> -n <namespace> --show-labels
 
# 4. Check resource requests are valid
kubectl get pod <pod-name> -o jsonpath='{.spec.containers[*].resources}'
 
# 5. Check subnet IPs
aws ec2 describe-subnets --query 'Subnets[*].{ID:SubnetId,Free:AvailableIpAddressCount}'

Resources

Free Tools
Free Tool

K8s Resource Calculator

Calculate CPU & memory requests/limits for your pods. Never get OOMKilled again.

Free Tool

CI/CD Pipeline Generator

Generate GitHub Actions, GitLab CI, or Jenkins pipelines for your stack instantly.

Back to all articles
Share:Tweet
Newsletter

Stay ahead of the curve

Get the latest DevOps, Kubernetes, AWS, and AI/ML guides delivered straight to your inbox. No spam — just practical engineering content.

Related Articles

Comments