Skip to content
This repository has been archived by the owner on Feb 2, 2024. It is now read-only.

Latest commit

 

History

History
65 lines (47 loc) · 2.5 KB

guide-one-click-request-sizing.md

File metadata and controls

65 lines (47 loc) · 2.5 KB

Guide: 1-click request sizing

⚠️ This feature is in a pre-release (alpha/beta) state. It has limitations. Please read the documentation carefully. ⚠️

1-click request sizing is a feature that will instantly update container resource requests in your cluster based on Kubecost's sizing recommendations. This allows you to instantly optimize resource allocation across your entire cluster, without fiddling with excessive YAML or arcane kubectl commands. You can easily eliminate resource over-allocation in your cluster, which paves the way for vast savings via cluster right-sizing and other optimizations.

Setup

Install Kubecost with cluster controller enabled, which is the only Kubecost component with write permission to your cluster.

# Make sure the Kubecost helm repo is set up! See kubecost.com/install

# This command will install Kubecost if you haven't already. You can use
# "--set clusterController.enabled=true" to get cluster controller running if 
# you already have Kubecost installed.
helm upgrade \
    -i \
    --create-namespace kubecost \
    kubecost/cost-analyzer \
    --set kubecostToken="aWljaGFlbEBrdQQY29zdljb203yadf98" \
    --set clusterController.enabled=true

Using 1-click request sizing

  1. Visit the request sizing page of your Kubecost installation

    kubectl port-forward \
        -n kubecost \
        service/kubecost-cost-analyzer \
        9090

    Then visit http://localhost:9090/request-sizing.html

  2. Modify the request sizing settings, like profile, window, and filters, until you have a set of recommendations you are ready to apply to your cluster.

  3. Click the "Automatically implement recommendations" button.

  4. Confirm!

Technical details

If you want to learn more about the APIs that power 1-click request sizing, including their limitations, see the API documentation.