Beta

Reports Cost now

Kubernetes cost

What each namespace and workload costs on your Kubernetes clusters, in AWS and on your own servers, and a month at this rate, by namespace and by cluster. The part of each node no pod held is priced too, as idle.

A pilot, on the hosted plan and on self-hosted installs: connect a cluster under Clusters, in Settings.

What it shows

The cost per hour of each namespace, with what its pods requested in CPU and memory and how much of that they used, the change over 7 days, and a month at this rate. Select a namespace and its workloads are listed the same way: deployments, stateful sets, jobs and daemon sets, each with its requested and used share. A trend over time draws the namespaces, bars put them side by side, a donut splits the cost by cluster and a heatmap shows the hour of day.

One row is (idle): the node capacity no pod reserved or used, priced like everything else. On an EKS cluster that is the idle capacity you pay EC2 for; on your own servers it is what the hardware costs you for nothing. A namespace whose pods use a quarter of what they request is paying for the other three quarters.

Questions it answers

  • What does each namespace cost a month on our EKS cluster, and which workloads make it up?
  • How much of our Kubernetes cost is idle capacity, and on which cluster?
  • Are the requests on this deployment far above what it uses?
  • What does a cluster on our own hardware cost, at what we pay for a core-hour and a GiB-hour?

What it reads

Every hour, for each Kubernetes cluster, one small agent pod installed in the cluster reads what each pod reserved (its CPU and memory requests), what it actually used, and which node it ran on. Names and labels, such as the app and the namespace, come from Kubernetes itself. The agent is open source, in the qloudpulse-public repository, and the only thing installed in a cluster: nothing is installed in your AWS accounts.

Each node's cost for the hour is split between the pods on it: a pod pays for the larger of what it reserved and what it used, with CPU and memory priced separately. A node in AWS costs what the AWS run rate says at list price. A node of a cluster on your own servers costs its cores and memory at the price per core-hour and per GiB-hour you set for that cluster; a changed price counts from the next hour read. Load balancers, volumes and the EKS control plane are not here; they are in the AWS run-rate report.

What to do with it

Start with the idle row. If it is a large share, the cluster has more nodes than its pods need: look at the node group's size, or at an autoscaler's settings. Then the namespaces whose used share is low: lower their requests, and the scheduler packs the nodes tighter. The AWS waste report carries the same idle cost as a line, next to what is idle elsewhere. For the pods behind a namespace over the last 12 hours, see Kubernetes cost per namespace.

Related reports

  • Kubernetes cost per namespace: What each namespace costs, and the CPU and memory its pods requested and used, over the last 12 hours, down to the pod and the node it ran on.
  • AWS waste: Money paid for what is not used, sorted by how easily it goes: unused, overprovisioned or stranded, by category and by service.
  • AWS run-rate cost: What AWS costs an hour now, by account, service and environment, priced from what is running rather than read from a bill that arrives a day late.
  • AWS savings recommendations: Changes that would cut the bill, priced per month, largest first, and what the changes made so far save.

Every plan has every report, hosted or self-hosted, and the hosted plan is free during early access: see pricing. Like every report, this one splits and filters by account, environment, service and resource, steps back 13 months, downloads as PDF and CSV and keeps the view in its link: what every report does. All reports.