Update EKS node groups timeout

When updating an Amazon EKS managed node group, there is a 15-minute timeout for draining Pods from the nodes being updated. If your application requires a longer delay and you cannot use the force update option, here are some best practices to consider:

Review and adjust your Pod Disruption Budgets (PDBs) to allow for a longer graceful termination period. PDBs control how many Pods can be unavailable during a voluntary disruption, such as a node update. By increasing the PDB limits, you can give your Pods more time to terminate gracefully.

Implement a pre-stop hook in your application's Pods to perform any necessary cleanup or graceful shutdown tasks before the Pod is terminated. This can help ensure that your application is properly prepared for termination within the allotted time.

Consider using a tool like Velero (formerly Ark) to back up and restore your application data and state during the node update process. This can help minimize downtime and ensure data consistency.

If the above options are not feasible, you may need to consider alternative approaches, such as: a. Scheduling maintenance windows for node updates during periods of lower traffic or usage. b. Temporarily scaling up your application to handle the disruption caused by the node updates. c. Breaking up your node group into smaller groups and updating them in phases, allowing more time for each phase.

Update a managed node group for your cluster - Amazon EKS