Skip to content
Platform Docs

Cost Optimization

Some hands-on tips & best-practices for optimizing cloud costs

Identifying Areas for Cost Optimization

AWS Cost Explorer & Right Sizing

Use the AWS Cost Explorer of your domain and filter by service. To drill down further, additionally group by usage type to identify areas to optimize.

Cost Explorer Example Diagram

The AWS Cost Calculator can be used to compare different setups and their costs. In the following example the costs of an database scale-down are compared:

Cost Calculator Example

Kubernets Right Sizing

Rightsizing information is available in Grafana in every cluster at https://grafana.eks-eu-central-1-private.<stage>.<domain>.carloop.dev/d/a87fb0d919ec0ea5f6543124e16c42a5/kubernetes-compute-resources-namespace-workloads?orgId=1&refresh=10s&var-datasource=default&var-cluster=&var-namespace=default&var-type=All

Search for the Dashboard "Kubernetes / Compute Resources / Namespace (Workloads)"

Select the default namespace

Check what are the requests vs the real usage, use a time selection which helps you to make a better decision ( e.g. at least daily, maybe weekly)

Adapt your requests in the base chart

MongoDB Atlas Database Costs

You can use the MongoDB Atlas Cost Explorer to get insights into your MongoDB Atlas database costs.

MongoDB Cost Explorer

For more information see also: https://www.mongodb.com/blog/post/view-analyze-your-monthly-mongodb-atlas-usage-cost-explorer

General Best Practices

  • use spot instances in Kubernetes (the default), do not use OnDemand instances
  • set an AWS Budget and check it regularly
  • regularly review database instance sizes
  • only forward logs to elasticsearch that you actually require
  • do not over-extend custom retention for elasticsearch indices

Out of Scope

The following things are out of scope for development teams, as the platform team already takes care about them:

  • procure reserved instances
  • use an AWS savings plan