EKS 1.31 extended support ends November 26, 2026, and 1.34 standard support ends December 2. What should I check before upgrading, and when does a new cluster make more sense than an in-place upgrade?
EKS 1.31 and 1.34 deadlines are coming up. How are you planning upgrades?
Answers
Start by finding every cluster and its version in each region. The AWS CLI commands list-clusters and describe-cluster can do that. You can also use describe-cluster-versions to check the support dates.
Then run list-insights for each cluster. Look for deprecated API usage and add-on or node version skew, and check those dependencies before scheduling the upgrade. After 1.31 extended support ends, AWS schedules a control-plane upgrade, so waiting means losing control over the timing.
Reply to this answer
For 1.34, this is a cost deadline rather than the end of extended support. Once standard support ends, the cluster rate goes from $0.10 to $0.60 per hour, roughly $365 more per month per cluster. Multiply that by your cluster count when deciding how much time extended support is worth.
Reply to this answer
In-place is the simpler path if the readiness checks are clean and you can test the upgrade and its dependencies first. A separate replacement cluster gives you another environment to validate before moving workloads, but it also means managing the migration. Neither option gets you out of checking add-ons and node compatibility.
Reply to this answer
Extended support buys time, but it doesn't remove the work. EKS upgrades still go through Kubernetes versions one at a time, so a cluster that's far behind needs a plan for each step.
There's a real tradeoff here. Longer support would give teams more breathing room, while regular upgrades keep security fixes and maintenance on the schedule. Either way, test new releases in a development environment and keep a lifecycle plan rather than treating the deadline as a one-off project.
Reply to this answer
Add your answer
Answers are reviewed before publication.
Related Discussions