https://catalogartifact.azureedge.net/publicartifacts/zoorik.coral-52018e26-68f9-4b54-b9bc-0cd42bab871b/image2_coralmarketplacelogo300216x216.png

Overview Ratings + reviews Details + support

Coral optimizes AKS clusters by right-sizing workloads and consolidating nodes.

Coral rebalances your AKS clusters so you pay for what your workloads actually use. It right-sizes container CPU and memory requests, empties the nodes you don't need, and moves work onto cheaper VM sizes. Built for platform, DevOps and FinOps teams running Azure Kubernetes Service.

What Coral does

Workload right-sizing: sets each container's CPU and memory requests from real usage (by default, CPU at the 80th percentile of the last 3 days and memory at the peak of the last 8 days, plus 10% headroom). Recommendations react within minutes to usage surges and out-of-memory kills.

Node optimization: works out how many nodes each pool really needs, checks on the live cluster that every pod fits elsewhere, then drains spare nodes gradually for removal.

Rebalancing: replaces the nodes you choose with cheaper VM sizes that still fit the work, through AKS Node Auto Provisioning.

Clear savings: every figure uses Azure's published list price and says where it comes from.

You stay in control

No workload is resized until you switch on "Automatically optimize" for a policy or a workload. Until then, Coral only recommends.

Node optimization starts at install. You can switch it off, or limit it to the node pools you choose.

Coral never creates machines itself. Scale-up stays with AKS Node Auto Provisioning or the Cluster Autoscaler.

Your manifests are never changed. Coral sizes pods as they start, and resizes running pods in place on Kubernetes 1.33 or later.

Pods and nodes marked do-not-disrupt are never moved.

Works with Microsoft

Azure Kubernetes Service (AKS), AKS Node Auto Provisioning and the Cluster Autoscaler.

Sign-in with Microsoft Entra ID.

Hosted on Microsoft Azure.

Requirements and limits

AKS on Kubernetes 1.30 or later, with at least one Linux node (amd64 or arm64).

cluster-admin access and the Azure CLI, to run the one-command connect with zoorikctl.

Windows nodes are listed and costed, but their workloads are not optimized.

How to get it

At a glance

https://catalogartifact.azureedge.net/publicartifacts/zoorik.coral-52018e26-68f9-4b54-b9bc-0cd42bab871b/image7_coralscreenshot01overview.png

https://catalogartifact.azureedge.net/publicartifacts/zoorik.coral-52018e26-68f9-4b54-b9bc-0cd42bab871b/image8_coralscreenshot02dashboard.png

https://catalogartifact.azureedge.net/publicartifacts/zoorik.coral-52018e26-68f9-4b54-b9bc-0cd42bab871b/image1_coralscreenshot04impact.png

https://catalogartifact.azureedge.net/publicartifacts/zoorik.coral-52018e26-68f9-4b54-b9bc-0cd42bab871b/image6_coralscreenshot05containerlivemigration.png

https://catalogartifact.azureedge.net/publicartifacts/zoorik.coral-52018e26-68f9-4b54-b9bc-0cd42bab871b/image4_coralscreenshot03workloadoptimization.png

Coral

by Zoorik