Phrontis Start free

Know what every namespace costs.

Cost per workload, idle capacity and the requests to change, at cloud list prices.

71.2% of capacity allocated 0 100%
Allocated Idle
Demo cluster, per month: six m5.large nodes at AWS list prices. Synthetic data.

Where $503.70 a month goes

Each container pays its request plus any usage above it. What a node has left over is idle, so namespaces plus idle add up to the cluster bill.

Demo cluster cost per month by namespace, with idle capacity
NamespaceMonthlyPodsShare
checkout$135.05526.8%
payments$81.55316.2%
data$71.18214.1%
monitoring$43.8028.7%
kube-system$26.8695.3%
idle$145.26028.8%

It reads. It writes only when you say so.

Install it in a production cluster and know exactly what it can touch.

What it reads

Nodes, pods, volumes, services, workloads and pod metrics from metrics-server. The agent sends its samples only to the API address you give it.

What it writes

Nothing, until you turn on automation. The node policy is applied by a command you run with your own kubeconfig, after it shows the diff and you type the context name.

How you stop it

Every change is in a journal with the spec before and after. The kill switch stops automation and leaves the cluster as it is.

From sign-in to the first report

  1. 1

    Sign in with GitHub

    Create an organisation and get a cluster key.

  2. 2

    Install the agent

    One helm install. The image is signed with cosign.

  3. 3

    Read the report

    Cost by namespace and workload, idle, and requests to change.

  4. 4

    Automate, if you want

    Dry-run first, then let Pro apply requests and node policy.

Free up to 500 vCPU and 3 clusters. No card.