💰 FinOps · Kubernetes · Multi-Cloud
Master Your Cloud & Kubernetes Bill
Allocate · Attribute · Rightsize · Chargeback
Pool-aware Kubernetes cost allocation (spot vs on-demand, GPU). Per-resource attribution across GCP, AWS and Azure via BigQuery, CUR Athena and FOCUS. Interactive dashboards, multi-threshold budgets, rightsizing recommendations and PDF chargeback — unified with the rest of your Ops, Security and AI stack.
Kubernetes Cost Allocation — Pool-Aware
Allocate cluster cost per namespace using max(requests, usage) on CPU and RAM over 14 days of Prometheus data, with a deterministic split between spot and on-demand capacity.
🟢 Spot vs On-Demand
Lifecycle mapping for Karpenter, EKS, GKE-spot and AKS. Per-namespace breakdown: spot, on_demand, unknown.
🏷️ Node Pool Breakdown
Sub-split by node pool (GKE, EKS, Karpenter, AKS) via labels: gke-nodepool, karpenter.sh/nodepool, agentpool.
🎮 GPU & Instance Type
Per instance type (node.kubernetes.io/instance-type) and GPU product (nvidia.com/gpu.product) — see exactly what your A100s and H100s cost.
Per-Resource Attribution — GCP · AWS · Azure
Pull cost data directly from the source of truth — no proxy, no re-billing — and break it down to the individual resource.
☁️ GCP
Detailed BigQuery billing export (billing_export_resource_table).
Per resource, per project, per service, per label.
🟧 AWS
CUR 2.0 via Athena (cur_athena_*).
Auto-detect resource_tags_* columns via information_schema, cached 24h per connector.
🔷 Azure
Cost Management blob export — FOCUS or actualcost (cost_export_*).
Per resource, per subscription, per resource group.
Unified model: CostAttribution(date, provider, service, cloud_resource_id, cloud_account_id, k8s_cluster, k8s_namespace, tags, amount_micros).
Interactive Dashboard — URL-Synced
Every filter lives in the URL. Share a view, bookmark a query, reopen exactly where you left off.
🎛️ Powerful filters
Provider, service, connector, tenant, K8s cluster, K8s namespace, tag key + value, date range. Pills UX, click-to-filter directly on ECharts and treemap, side panel for per-resource drill-down.
📊 Server-side breakdowns
group_by computed in the API: service, provider, account, connector, tenant, cluster, namespace. Mixed-currency detection with explicit warning — no silent FX conversion.
Budgets & Multi-Threshold Alerts
🎯 Scoped budgets
Monthly or quarterly periods. Filter by provider, service, K8s cluster, namespace, tags. Multiple thresholds per budget (50 %, 80 %, 100 %) with linear MTD forecast.
🔔 Multi-channel alerts
Push to Slack, Discord, Teams, Telegram, email or webhook. Daily evaluation (finops.evaluate_budgets 06:30 UTC). Dashboard banner whenever your active filters are a subset of a budget scope.
Rightsizing Recommendations
Weekly scan (Sunday 06:00 UTC) detects waste and pushes actionable recommendations with estimated savings.
🧮 CPU & RAM
Over- and under-provisioned workloads detected from 14d of Prometheus data, with severity (low / medium / high).
📦 BestEffort & Unused PVC
Catch pods running without requests (BestEffort QoS) and persistent volume claims with no I/O — silent budget killers.
🧊 Snooze & Auto-Resolve
Snooze, dismiss or resolve. Findings auto-resolve when they disappear from the next scan — no stale backlog.
PDF Chargeback Export
One endpoint, one PDF: GET /finops/chargeback.pdf?<filters>. Throttled (5 req/h/org) to keep things sane.
- ✓ Cover page with active filters
- ✓ Totals + daily bar chart
- ✓ Breakdowns per service and tenant
- ✓ Top resources by cost
- ✓ Server-side rendering with reportlab — no client-side surprises
Ready to forward to Finance, attach to a Jira ticket, or archive as evidence for budget reviews.
Frequently Asked Questions
How does Argonix differ from Cloudability, Vantage or CAST AI?
Argonix unifies FinOps with the rest of your Ops, Security and AI Ops platform — the same connectors that power monitoring and CSPM power cost attribution. No extra agent, no duplicated inventory. Plus, the Argos agent can reason about cost questions (« why did GKE spend spike yesterday? ») using the live data.
Do you do FX conversion on multi-currency clouds?
No. We surface a mixed_currency warning with the list of currencies whenever the active filters span multiple billing currencies. Silent FX hides real numbers from Finance — we refuse to do it.
What data sources are required?
For GCP: detailed BigQuery billing export. For AWS: CUR 2.0 with Athena. For Azure: Cost Management blob export (FOCUS or actualcost). For Kubernetes: a Prometheus connector with 14 days of kube-state-metrics / cAdvisor data.
Can budgets target specific teams or services?
Yes. Budget filters use the same dimensions as the dashboard: provider, service, cluster, namespace, tag key + value. Combine them to scope a budget to one team, one product, one environment — or all of them.
Is FinOps included in the Pro plan?
Core FinOps features (dashboard, attribution, K8s allocation) are included in Pro. Advanced features (large data volumes, dedicated BigQuery / Athena performance tier, chargeback at scale) are part of Enterprise. Talk to us for sizing.
Explore More
Stop Guessing What Your Cloud Costs
Spot vs on-demand · GPU breakdown · per-resource attribution · budgets · rightsizing · PDF chargeback. All in one sovereign platform.