Troubleshooting common issues on Fargate

The most common issue related to capacity is that occasionally, there will not be enough platform resources, and/or the CPU/RAM combination you want will not be supported. This results in the Pod status always being PENDING. If it is a platform issue, simply waiting and trying later (after 15/20 minutes) may resolve the issue; otherwise, adjust your Pod spec to support a different CPU/RAM combination.

If your Fargate nodes are shown as Not Ready when you run the $ kubectl get nodes command, ensure that the execution role they are using is also configured in the aws-auth ConfigMap, an example of which is shown as follows:

mapRoles: |
    - groups:
      - system:bootstrappers
      - system:nodes
      - system:node-proxier
      rolearn: <Pod_execution_role_ARN>
     ...