This section captures best practices with Amazon EKS Clusters that will help customers deploy and operate reliable and resilient EKS infrastructure.
Ensure that you select the latest version of k8s for your EKS Cluster. This will ensure that you do not have to deal with upgrades immediately.
Pods per Instance Type¶
Ensure that you select appropriate instance types for your worker nodes. Amazon EKS's limits for "number of pods per instance type" are documented here
AWS Node Termination Handler¶
Ensure that you deploy AWS's node termination handler to your EKS clusters. On Amazon EKS clusters that are provisioned and managed by Rafay, AWS's node termination handler is automatically deployed to ensure interruptions are handled gracefully.
IP Addresses in Subnet¶
Ensure you provide sufficient IP addresses in your subnet. Pod scheduling may not be possible if you run out of IP addresses in your subnet. More details in Troubleshooting EKS CNI Guide