Evaluation
Evaluate KubeOn on your own bill
An evaluation runs on your clusters and your billing data, with an engineer from our team. Here is what to prepare, what you see and when, and what you keep at the end.
How it starts
Three steps before the first numbers
- 01
Agree the scope
On a 30-minute call, choose the clusters, billing sources and teams to include, and what the evaluation should show. Scope, length and access are agreed before it starts.
- 02
Connect, read-only
An engineer from our team helps you install the agent and set up read-only billing access. Every permission is published on the security page.
- 03
Review together
Go through the reconciliation, ownership coverage, unallocated cost and the first change proposals, then decide on next steps.
Before you start
What to have ready
One cluster and read access to its bill are enough to start. Each billing source needs one export and one read-only identity.
| Every cluster | Cluster-admin once, to create the agent's read-only ClusterRole. Outbound HTTPS to KubeOn Cloud, or a bucket you own for snapshots. A Prometheus-compatible endpoint is optional. |
|---|---|
| AWS | A CUR 2.0 Data Export with resource IDs and split cost allocation data, cataloged in AWS Glue, and the read-only billing role deployed with CloudFormation. A new export can take up to a day to deliver its first file. |
| Azure | A Cost Management export, amortized cost or FOCUS, to a storage account, read with a managed identity. |
| Google Cloud | The Cloud Billing detailed export in BigQuery, read with a service account that has BigQuery Data Viewer and BigQuery Job User. |
| Data centers | A rate card per node pool or label, or hardware cost over a depreciation period plus power, space and support. |
What you see, and when
From install to the first proposals
01
Within the first hour
The cluster, its nodes, namespaces and workloads appear in Diagnostics. Allocated cost per namespace, team and label follows as soon as billing data is in.02
The same day
Once billing data is available, 180 days of history are read and allocated, plus 30 days of per-pod split cost history where AWS split cost allocation data exists. A brand-new AWS export can take up to a day to deliver its first file. The allocated history is kept; the raw billing rows leave the hub after 90 days.03
Within 30 minutes of the first allocation
Savings are computed and the first change proposals appear, starting with findings that need no usage history, such as released volumes that still bill.04
After 7 days of usage
Container requests are right-sized from p95 usage. With Prometheus history connected, this starts on the first day.05
After 14 days
Node group sizing, abandoned workloads and over-provisioned volumes, which each need 14 days of data.
What you keep
What the evaluation leaves you with
An engineer from our team reviews the first allocations and savings with you.
A reconciliation per cluster
Billed cost against allocated cost, ownership coverage, and every unmatched resource listed by ID.
Cost by team and label
Allocation by team, namespace and any label, with shared and idle cost in their own columns.
Savings with proposals
Each saving priced against your bill, with evidence, a risk level and a patch your team can review.
A data health report
The 10 checks for every cluster, and the gaps to fix before you roll out across the fleet.
Your data
Read-only throughout, and yours at the end
- The agent has get, list and watch only, and never reads Secrets or ConfigMaps.
- Billing is read through a read-only role, managed identity or service account.
- On KubeOn Cloud, evaluation data is deleted when the evaluation ends unless you continue.
- On a self-hosted hub, nothing leaves your account.
FAQ
Questions before an evaluation
Long enough for the findings you care about. Allocation and reconciliation are ready on the first day, container rightsizing needs 7 days of usage, and node group sizing needs 14.
Request an evaluation
Start with a scoping call
Tell us where your clusters run and which ones you want to include. We reply within one business day to set up the 30-minute call.
Not ready to connect a cluster? Book a demo