Applying Recommendations

When you identify a workload to optimize, click on it to open the Workload Details page. This page shows the current resource configuration, the recommended changes, and the expected impact on cost and reliability. On the top left of the screen you can find the Expected Impact of the configuration and the "Apply Now" button, which provides instructions to apply the recommended configuration for the workload on your system.

Ways to apply a recommendation

Akamas Insights provides three ways to apply recommendations.

Apply Configuration Drawer

1. kubectl Command

The fastest method for applying changes directly to running resources. Copy the provided kubectl patch command, adapt it to your specific environment, and execute it in your terminal. The command applies changes immediately. This method is useful for quickly deploying configurations.

2. YAML File

Download the recommended configuration as a YAML snippet and merge it into your deployment files. This approach fits GitOps workflows and ensures changes are persisted in source control. After updating your YAML files, apply them with kubectl apply -f <your-file.yaml> or your preferred tool of choice, such as a Helm or Kustomize.

3. GitOps Integration

Akamas Insights can create a merge request with the recommended changes directly in your repository by integrating with your provider of choice such as GitLab, GitHub, or Bitbucket. This automates the process of updating your infrastructure-as-code and provides a review workflow before changes are applied.

HPA Recommendations

For workloads managed by a Horizontal Pod Autoscaler, the apply flow includes HPA-specific configuration in addition to container resource changes. The HPA section appears first in the apply drawer and includes:

• Recommended maxReplicas: The optimal upper bound for replica scaling based on observed peak demand;

• Recommended minReplicas: The optimal lower bound based on baseline demand.

HPA changes target the HPA resource itself (not the Deployment), so the generated kubectl command and YAML snippet will patch the HPA object rather than the workload containers. Both HPA and workload configuration changes can be applied together in a single operation.