Struggling with Helm and Looking for Better Templating Options for K8s

0
1
Asked By CuriousCat123 On

Hey folks! I'm new to Kubernetes (K8s) and I've been having a rough time with Helm. I've decided to only use Helm for distribution later on if I share my projects (which are mainly for internal use). So, I'm looking to refine my development workflow with a better templating language.

Currently, I'm pointing Skaffold at a directory where I'm rendering YAML manifests using a templating tool. I've tried out CUE and KCL but I'm torn between them. I hear that KCL is simpler yet more powerful than CUE, but the development activity on KCL seems quite low. I'm not sure if it's worth investing time into KCL or if I should just go with CUE. What do you think? Is CUE the better option?

3 Answers

Answered By TechExplorer99 On

I totally get your frustration! I've heard similar things about Kustomize being unintuitive, although it’s part of the kubectl package. You might expect it to be the go-to choice just because it’s officially supported, but many of these tools can feel pretty clunky. Honestly, it’s worth revisiting Kustomize if you haven't given it a fair shot yet, though most templating languages do have a learning curve.

Answered By DevGuru88 On

I think most people go through this stage! I've tried Helm, Kustomize, ytt, jsonnet, and even some scripting with bash and Python. I don't hate Helm as much as others—it's not perfect, but it works. A helpful tip: keep your business logic and templating logic separate. Make changes and review the output immediately. Sometimes the issue lies in how the Helm charts are structured. People often overlook using helpers and managing variables effectively.

Answered By CodeCrafter45 On

I can relate to what you’re saying. I found popular tools like Helm and Kustomize lacking in clarity and flexibility. After struggling for a couple of days with Kustomize and getting zero error messages as things went wrong, I switched to ytt and kapp, which felt like a better compromise for internal deployment. They provide clearer error messages and support Python, which is pretty handy. CUE and KCL are definitely worth checking out, though!

Related Questions

LEAVE A REPLY

Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.