What’s the best way to handle over 1,000 concurrent Helm installs?

0
7
Asked By CloudSailor88 On

I'm managing a whole bunch of Helm releases (we're talking 1,000+!), and while I'm currently using Helmfile, I find it a bit slow and quite the memory consumer. I've heard mixed things about its concurrency abilities and it feels like it could do better with utilizing multiple cores, although I plan to look into that more. I'm curious if anyone has discovered any nifty tricks or if I should just run several Helmfile instances at once. Would love to hear your thoughts!

3 Answers

Answered By KustomizeKing On

Have you compared Helmfile with Kustomize? Kustomize can utilize Helm charts but works differently by rendering them first. It might be worth exploring to see if it handles your scale better. Also, if you’re considering ArgoCD, you could organize your deployments with the App of Apps pattern if everything’s in one cluster.

Answered By FluxMasterX On

FluxCD is definitely worth checking out! I found it easy to manage Helm releases, especially when scaling up. You might want to look at their guides for more insights on efficiency with large deployments.

HelmHero7 -

Totally agree! Flux seems like a practical choice for Helm releases. We haven't hit the same scale yet, but it's been smooth sailing so far.

Answered By TechieTribe42 On

Have you looked into using ArgoCD? I’ve set it up with a management cluster before and it worked pretty well to manage applications across multiple clusters. You can use an ApplicationSet to deploy your charts, and it handles larger setups quite effectively.

DeployDiva99 -

I've been considering ArgoCD as well, but I thought I’d seen something about limitations with ApplicationSets for larger numbers. How has your experience been with Argo handling numerous Helm releases?

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.