Onboarding Walkthrough

Once your agent is running and data is flowing, this guide explains what you’re looking at in the CostPilot dashboard and what actions to take first.

What you’ll see on first login

Tip

New accounts automatically enable demo data — a fictional cluster called “Demo Production Cluster” that lets you explore all CostPilot features before your real data appears. You can disable this in Settings → Features.

The dashboard is divided into three main areas:

1. The summary bar

At the top of the Dashboard, you’ll see four headline numbers:

  • Total cost — Your cluster spend for the selected period (default: last 7 days)
  • Efficiency score — A percentage representing how well your cluster resources are being used. Higher is better. See Efficiency Scoring for how this is calculated.
  • Idle cost — How much you’re paying for unused capacity. This is your biggest lever for savings.
  • Pod count — The number of distinct pods CostPilot has observed over the period.

2. The cost breakdown

Below the summary bar, you’ll see cost broken down by dimension. By default this shows costs by namespace, but you can switch to workload, node, or any of your Kubernetes labels via the Cost Explorer.

Each row shows:

  • Absolute cost for the period
  • Percentage share of total spend
  • An efficiency indicator (colour-coded)

The trends chart shows daily cost over the selected time period. Look for:

  • Sudden spikes — Often a new deployment or autoscaler event
  • Steady upward trends — Growth in workload or resource requests
  • Flat idle cost — Capacity that isn’t being used but is still being paid for

First actions after connecting your cluster

Check your efficiency score

An efficiency score below 50% typically indicates significant overprovisioning. Click into Cost Explorer → Namespace to see which namespaces are the least efficient.

Review idle costs

Open the Dashboard and look at the Idle cost figure. If this is more than 15–20% of total spend, you likely have overprovisioned nodes or workloads with very low resource usage relative to their requests. See Understanding Idle Costs for how to reduce this.

Configure label dimensions

If your teams use Kubernetes labels to identify workloads (e.g. team, environment, cost-centre), set these up in Settings → Labels. This unlocks the most powerful cost attribution features in CostPilot — you’ll be able to see how much each team is spending in real time.

Enable insights

Once you’ve collected enough data (typically after 24 hours), CostPilot will generate automated recommendations. These appear under Insights and are ranked by potential monthly savings. Basic plan accounts run 4 analysers (~19 insight types); Pro and Max plan accounts run all 11 analysers (40+ insight types).

Set up alerts

If you want to be notified when costs exceed a threshold, configure alerts in Settings → Alerts. Alerts are a paid feature and support Slack, email, and webhook delivery.

SectionWhat it does
DashboardSummary metrics, efficiency, trends
Cost ExplorerDimensional drill-down with filters
AnalysisVelocity, breakdown, signals, and efficiency trends
InsightsAutomated recommendations
NodesFleet health, node utilisation, cluster and node management
PerformanceResource provisioning and right-sizing analysis
NotificationsReal-time operational events
SettingsLabels, alerts, billing, team, and feature configuration

Next: Explore the Dashboard guide for a detailed breakdown of every metric shown on the main page.