Live metrics processing

Know what your
Cluster costs

Real-time cost tracking for your entire Kubernetes infrastructure. Track spending by cluster, namespace, workload, or any label you choose.

45%
Avg cost reduction
<50ms
Query latency (p95)
100%
Cost visibility
Multi-cloud support 14-day free trial No credit card required
COST BREAKDOWN / LAST 7 DAYS
production
47 workloads
$5,894
↓ 23%
staging
12 workloads
$868
↑ 12%
development
8 workloads
$342
→ 0%
Total Cost $7,104
Idle Resources $1,240 (17%)
CRITICAL
2h ago
production/api-service oversized
Reduce CPU from 2000m to 500m
Save $890/month
TOP COST LABELS
team 12 values
environment 3 values
app 47 values
PLATFORM FEATURES

Built for Platform Teams

Everything you need to understand, track, and optimise your Kubernetes costs.

Real-Time Cost Tracking

Track costs per namespace, workload, cluster, and node with configurable collection intervals. See exactly where your money is going as metrics flow in.

→ Know your costs before the monthly bill arrives
Pod-level attribution Idle cost tracking Multi-cluster view

Automated Insights

Automatic detection of cost anomalies, oversized resources, and optimisation opportunities. Get actionable recommendations with estimated savings based on actual resource usage patterns.

→ Proactive cost optimisation without manual analysis
Auto-detect waste Severity classification Savings estimates

Smart Label Optimisation

As metrics accumulate, automatic cardinality analysis identifies which labels actually matter for cost tracking. Pin important labels, ignore the noise.

→ Focus on meaningful cost dimensions
Cardinality analysis Auto-indexing Pin/unpin labels

Flexible Cost Allocation

Allocate costs by any label, namespace, or custom dimension. Configure CPU/memory ratios to match your workload patterns. Set custom pricing per cluster or provider.

→ Accurate chargebacks without spreadsheets
Label-based splits Configurable ratios Team budgets

Performance Correlation

See CPU throttling, memory pressure, and resource usage alongside costs. Understand if cost spikes are from growth or inefficiency.

→ Connect performance issues to cost impact
Resource metrics Trend analysis Anomaly detection

Multi-Cluster Management

Single dashboard for all your clusters across AWS, GCP, Azure, and on-prem. Compare costs, track trends, and identify outliers.

→ Complete visibility across your infrastructure
Unlimited clusters Cross-cloud support Unified metrics

How CostPilot Compares

CLOUD PROVIDER TOOLS
• Monthly billing only
• No Kubernetes context
• Limited label support
TRADITIONAL SOLUTIONS
• Per-node pricing
• Manual configuration
• Slow data updates
COSTPILOT
• Real-time metrics
• Dashboard-driven config
• Automatic optimisation
BEFORE OPTIMIZATION
pod-template-hash 28,473 values
controller-revision-hash 12,891 values
pod-name 8,234 values
team 12 values
environment 3 values
High cardinality labels waste storage and slow queries
AFTER OPTIMIZATION
team
12 values PINNED
environment
3 values PINNED
app
47 values AUTO
✓ Optimized for cost tracking • Fast queries • Efficient storage
SMART LABEL OPTIMISATION

Stop tracking every label

Not all labels matter for cost tracking. CostPilot automatically analyses label cardinality as metrics accumulate and identifies which labels actually provide useful cost insights.

Automatic Cardinality Analysis

As metrics accumulate, the system learns which labels have low cardinality (useful for grouping) vs. high cardinality (wasteful for storage).

Pin Important Labels

Pin labels like "team" or "environment" to always track them. The system automatically indexes pinned labels for fast queries.

Auto-Optimize Storage

High-cardinality labels like pod-template-hash are automatically excluded from cost tracking, saving storage and improving query performance.

AUTOMATED INSIGHTS

Stop hunting for savings.
Let insights find you.

As metrics accumulate, automated analysis detects anomalies, identifies waste, and generates actionable recommendations with estimated savings.

1

Metrics Accumulate

As your agent collects metrics, usage patterns emerge. The system tracks CPU, memory, and storage consumption across all workloads.

2

Thresholds Trigger Analysis

When metrics cross thresholds (idle workload, oversized pod, cost spike), the system automatically generates an insight with severity classification.

3

Get Recommendations

Each insight includes a specific recommendation, estimated monthly savings, and impact assessment. Track resolution status in your dashboard.

Example Insights

CRITICAL Oversized Resources Detected
2 hours ago
ISSUE
production/api-service using 15% of requested CPU
RECOMMENDATION
Reduce CPU request from 2000m to 500m
POTENTIAL SAVINGS
$890/month
IMPACT
High cost impact, low risk
WARNING Idle Workload
2 hours ago
ISSUE
staging/old-worker has zero traffic for 30+ days
RECOMMENDATION
Consider decommissioning or scaling to zero
POTENTIAL SAVINGS
$124/month
IMPACT
Quick win, zero downtime
INFO Spot Instance Opportunity
2 hours ago
ISSUE
dev-cluster workloads suitable for spot instances
RECOMMENDATION
Migrate 12 workloads to spot node pool
POTENTIAL SAVINGS
$340/month
IMPACT
60% cost reduction on dev workloads
INSIGHT CATEGORIES
→ Oversized resources (CPU/memory waste)
→ Idle workloads (zero traffic)
→ Cost anomalies (sudden spikes)
→ Performance issues (throttling, OOM)
→ Spot instance opportunities
→ Storage optimisation
STATUS TRACKING
Active - Newly detected
Acknowledged - Team is aware
Dismissed - Not applicable
Resolved - Action taken
Auto-resolved - Fixed automatically
HOW IT WORKS

From Zero to Full Cost Visibility

Deploy the agent, let metrics flow, and get actionable insights—all configured from your dashboard.

1

Deploy the Agent

# Add Helm repo and install agent
$ helm repo add costpilot https://charts.costpilot.dev
$ kubectl create namespace costpilot-system
$ helm install costpilot-agent costpilot/agent \
--namespace costpilot-system \
--set clusterId=prod-us-east \
--set clusterApiKey=cp_live_...
✓ Agent deployed, collecting metrics
Configure collection interval and pricing in dashboard →

Get your cluster ID and API key from the dashboard. The agent runs with minimal resources and collects metrics at intervals you configure.

2

Metrics Flow In

Pod metrics 847 pods
Node metrics 12 nodes
Collection interval Configurable

Agent scrapes metrics-server, batches efficiently, and ships to our ingestion pipeline. All collection settings controlled from your dashboard.

3

Costs Calculated

CPU cost $2,847
Memory cost $1,940
Storage cost $573

Allocator calculates costs using pricing models you configure per cluster. Supports on-demand, spot, reserved, and custom rates.

4

Automatic Analysis

As metrics accumulate, insights and label optimisation kick in

LABEL ANALYSIS
→ Cardinality analysis running
→ Found 3 high-value labels
→ Auto-indexed: team, environment, app
→ Ignored: pod-template-hash (28k values)
INSIGHTS GENERATED
→ 3 CRITICAL issues detected
→ 12 WARNING opportunities found
→ 42 INFO recommendations
→ Est. savings: $11,240/month
5

Configure Everything

Collection Interval
15s → 30s → 60s
Pricing Model
On-demand / Spot / Reserved
CPU/Memory Ratio
40% / 50% / 60%

All agent configuration happens in the dashboard. No YAML editing, no cluster access needed.

6

Track & Optimise

REAL-TIME DASHBOARD
Total cost (7d) $7,104
Idle resources $1,240 (17%)
Active insights 57
Potential savings $11,240/mo

View costs by cluster, namespace, workload, or any label. Export data, set alerts, track optimisation progress.

Ready to see your costs clearly?

Start Free Trial
14-day free trial • No credit card • Deploy in minutes
REAL-WORLD SCENARIOS

Built for Your Challenges

These aren't hypotheticals—they're the problems platform teams solve every day with CostPilot.

Multi-Tenant Cost Attribution

PROBLEM:

Running a platform serving multiple teams or customers

SOLUTION:

Automatic cost breakdown by namespace, label, or custom dimensions

OUTCOME:

Accurate chargebacks without manual tracking or spreadsheets

BEFORE
$180K/mo → "Platform Costs"
AFTER
Team A: $72K | Team B: $54K | Team C: $36K | Shared: $18K

Spot Instance Optimisation

PROBLEM:

Paying on-demand prices for workloads that could run on spot

SOLUTION:

Real-time tracking of node pricing types with optimisation recommendations

OUTCOME:

40-70% cost reduction on suitable workloads

BEFORE
100% on-demand nodes
AFTER
70% spot, 20% reserved, 10% on-demand | 58% savings

Resource Right-Sizing

PROBLEM:

Over-provisioned pods wasting money, or under-provisioned causing throttling

SOLUTION:

Actual usage vs. requested resources with performance correlation

OUTCOME:

Eliminate waste while maintaining SLOs

BEFORE
api-service: 2 CPU req, 0.3 CPU avg → $420/mo wasted
AFTER
Rightsized to 0.5 CPU req | Saved $420/mo | 0 throttle events

Budget Management

PROBLEM:

No way to track team spending against allocated budgets

SOLUTION:

Label-based cost tracking with alert thresholds and trend analysis

OUTCOME:

Teams stay within budget, finance gets accurate forecasts

BEFORE
Quarterly surprise: Team X spent 180% of budget
AFTER
Alert at 80% threshold → proactive optimisation → on budget
45%
Average cost reduction
2hrs
Average time to first insight
100%
Cost visibility
0
Agent maintenance required
PRICING

Simple, Transparent Pricing

No per-pod pricing, no surprise overage fees. Pay for what you need, scale as you grow.

Basic

$ 199 /month

For small teams getting started with cost visibility

  • Up to 3 clusters
  • Real-time cost tracking
  • Basic label allocation
  • 7-day data retention
  • Email support
  • Custom pricing models
  • Advanced insights
  • SSO
Start Free Trial
MOST POPULAR

Professional

$ 599 /month

For growing platform teams needing advanced features

  • Up to 20 clusters
  • Real-time cost tracking
  • Advanced label allocation
  • 90-day data retention
  • Priority support
  • Custom pricing models
  • Advanced insights & alerts
  • SSO (SAML/OIDC)
Start Free Trial

Enterprise

Custom

For large organisations with complex requirements

  • Unlimited clusters
  • Real-time cost tracking
  • Advanced label allocation
  • Custom data retention
  • Dedicated support + SLA
  • Custom pricing models
  • Advanced insights & alerts
  • SSO + SCIM provisioning
  • Self-hosted option
  • Custom integrations
Contact Sales
ALL PLANS INCLUDE
  • → 14-day free trial, no credit card required
  • → Real-time metrics processing
  • → Multi-cluster support
  • → API access
  • → Unlimited users
  • → Community Slack support
VOLUME DISCOUNTS
50+ clusters: 15% off
100+ clusters: 25% off
Annual contract: 20% off
Contact sales for custom pricing at scale

Have questions about pricing or features?

Contact our team →

Stop guessing.
Start measuring.

Join platform teams at leading companies who've gained complete visibility into their Kubernetes costs.

✓ 14-day free trial ✓ No credit card ✓ 5-minute setup
CostPilot

Real-time Kubernetes cost monitoring and optimisation platform built for platform engineers.

© 2026 CostPilot. Built for platform engineers.