How would you manage dev on EC2 and production on EKS without spending a lot more?

November 16, 2026 5 answers
Asked by Faye

Dev runs on EC2, while production uses EKS for deployments and scaling. The difference makes production changes hard to test, but a full-time UAT cluster would add cost. What would you change?

Answers

_h47~buttongrommet.2

I would aim for a smaller, disposable Kubernetes environment rather than a full-sized copy of production. Keep the deployment configuration and Kubernetes features close to prod, then bring it up when you need to test a release and tear it down afterward.

That gives you somewhere to catch deployment differences without paying for a full UAT environment all the time. It still won't be equivalent to production for load testing unless you provision enough capacity for the test.

6b~portgauge_391:

Automating the setup and teardown matters here. Otherwise the savings can get eaten up by the time spent rebuilding the environment.

Reply to this answer
Veya

A separate EKS test cluster with a node group that can scale down to zero is another option. Scale the nodes up for testing and back down afterward. You'll still need to account for the cluster and any other resources that remain running, so compare the whole bill rather than just node hours.

Spot capacity could reduce compute costs, though interruptions make it a less predictable choice for some tests.

Reply to this answer
Tavrel-9w

You could put a staging namespace in the existing production cluster to avoid another cluster. That's a cost tradeoff, not equivalent isolation. A problem in one namespace can affect the cluster as a whole, so decide whether that risk is acceptable before sharing it.

Reply to this answer
_m37!cachewrench.8

Containerizing the app consistently across EC2 and EKS can help with application-level parity, but it won't test the Kubernetes deployment and scaling behavior you're relying on in production. A small self-managed Kubernetes setup in dev is possible too, but factor in the time spent operating it, not just the infrastructure bill.

Faye:

Production uses Kubernetes for deployments and scaling, so matching the container alone wouldn't cover all the changes I need to test.

Reply to this answer
w.619!bytefunnel

Don't overlook the data side of this. A test environment can get expensive if it carries more data and supporting resources than its tests need. Decide what each test needs to load before copying a production-sized setup.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions