Skip to main content
Cloud Cost gives you an estimated view of what your clusters cost to run, without wiring up a separate cost tool. Ankra combines the live node and volume inventory it already syncs from your cluster with public cloud pricing to produce per-cluster and fleet-wide estimates, broken down by category and allocated back to namespaces and stacks.
All figures are estimates, derived from on-demand and spot pricing for the resources Ankra can see. They are designed for trend-watching, right-sizing, and chargeback - not for reconciling a cloud invoice to the cent.

How it works

To estimate cost for a cluster, Ankra needs three things:
  1. A cloud credential for the cluster’s provider (for example AWS or GCP), so it can resolve instance and volume pricing. See Credentials.
  2. The cluster to be online, so the agent can sync the current nodes and volumes.
  3. Pricing for the region and instance types in use.
A background inventory sync keeps the node and volume picture current per provider; cost is recomputed from that inventory plus your pricing settings.

Readiness states

The per-cluster cost view reports a readiness state so you always know why an estimate is or isn’t available:
StateMeaning
readyA cost estimate is available
no_credentialNo cloud credential is attached for this provider
unsupported_providerThe cluster’s provider isn’t supported for cost estimation
cluster_offlineThe cluster is offline, so inventory can’t be synced
awaiting_nodesWaiting for the first node inventory sync
awaiting_pricingNodes are known, pricing is still resolving
estimate_pendingInventory and pricing are present; the first estimate is being computed

Per-cluster cost

Open a cluster and select the Cost view to see its estimate. It includes:
  • Summary - hourly cost, estimated monthly cost, month-to-date, and a projected month-end total, with a confidence level.
  • Breakdown - compute split into on-demand and spot, plus storage, network, and control-plane cost, along with idle and unallocated portions.
  • Coverage - how many nodes and volumes were priced. If some couldn’t be priced, the estimate is flagged as coverage_incomplete so you know it understates the true figure.
  • Per-namespace allocation - cost allocated to each namespace (and the owning stack, where known) by CPU and memory share, so you can attribute spend to teams or workloads.
  • Trend - a daily series of the monthly cost estimate so you can spot changes over time.
GET /org/clusters/{cluster_id}/cost

Fleet summary

The organisation-wide rollup aggregates every cluster’s estimate into a single number, with breakdowns by provider and a list of your top-spending clusters:
GET /org/cloud-cost/summary
It returns estimated monthly cost, month-to-date, and projected month-end totals - in your configured display currency - both overall, per provider, and for the top clusters. This is the same rollup shown on the Dashboard.

Settings

Organisation admins can tune how cost is displayed and estimated under cost settings:
SettingDescription
currencyDisplay currency: usd, eur, or gbp
effective_discount_pctA flat discount (0-100) applied to list pricing, to approximate committed-use or enterprise agreements
include_network_egress_estimateWhether to add an estimated network egress component
GET /org/cloud-cost/settings
PUT /org/cloud-cost/settings   # organisation admins only
Only organisation admins can change cost settings. The discount must be between 0 and 100, and the currency must be one of usd, eur, or gbp.

Prerequisites checklist

1

Attach a cloud credential

Add an AWS or GCP credential for the cluster’s provider under Credentials.
2

Bring the cluster online

Ensure the agent is connected so node and volume inventory can sync.
3

Set your currency and discount

As an org admin, set the display currency and any committed-use discount in cost settings.
4

Read the estimate

Open the cluster’s Cost view, or the fleet rollup on the Dashboard.