You get a single view of cost and security across clouds by putting every account under one tagging and ownership scheme, exporting each cloud’s billing data in the FinOps Foundation FOCUS format and each cloud’s security findings into one place, and joining them on resource IDs and owner tags. Use a FinOps platform plus a CNAPP if you want to buy it, a data warehouse if you want to build it, and reduce how much there is to join by running workloads on fewer, consistently governed platforms.

Many organizations run workloads on two or more cloud providers, and each provider keeps its own billing and security data. Finance sees cost in separate billing consoles, and security sees findings in separate posture dashboards. Few teams can say how much they spend on resources with critical security findings, or who owns those resources. This guide covers the approaches available today, the data each cloud exports, and a six-step pipeline that joins cost and security on the same resource.

Disclosure: Control Plane publishes this guide.

TL;DR

A single view comes from joining billing exports and security findings on resource ID and owner tag.

  • The shared foundation. One tagging standard, one asset inventory, and one identity model across every cloud account are prerequisites. Without them, no tool can produce accurate, owner-attributed data.
  • Buy path. Pair a FinOps platform (Vantage, CloudZero, Finout, or nOps) with a CNAPP (Wiz, Prisma Cloud, Orca, or Microsoft Defender for Cloud) and join their outputs on resource ID and owner tag.
  • Build path. Export billing in FOCUS format and security findings in OCSF or each tool’s native export to a data warehouse like BigQuery or Snowflake, then build BI dashboards on top.
  • Joined metrics worth tracking. Spend on resources with critical findings, cost of untagged resources, owner with the most findings per dollar spent, and mean time to remediate findings on high-spend resources.
  • Where Control Plane fits. Workloads on Control Plane run from one definition with the same access policies and security controls on AWS, GCP, Azure, and Oracle Cloud Infrastructure (OCI). Their spend lands in one Cost & Usage view, which shrinks what the pipeline has to join to the cloud services that run outside Control Plane.

How can I get a single view of cost and security across multiple clouds?

The answer has three layers: a shared foundation, a tooling strategy, and a data pipeline that joins cost and security on the same resource.

The shared foundation is the hardest part. Every cloud account, subscription, and project must follow one tagging and ownership scheme. Tags like owner, team, environment, and service must be enforced by policy at creation time. Without consistent tags, a FinOps tool can report what an AWS account spent and a CNAPP can report its critical findings, but neither can tell you which team owns the problem or which specific resources are both expensive and insecure.

You also need one asset inventory keyed by resource ID. Each cloud has its own resource identifier format. AWS uses ARNs, Azure uses resource IDs, Google Cloud uses full resource names (for example, //compute.googleapis.com/projects/...), and OCI uses OCIDs. Your inventory must normalize these into a single lookup table so that billing records and security findings can be joined on the same key.

Finally, you need one identity model. Cross-cloud permissions are a major source of both cost waste and security risk. Excessive permissions lead to resources being provisioned without approval, and orphaned service accounts create attack surfaces that nobody monitors.

Once the foundation is in place, you choose a tooling strategy, and each common approach covers a different slice of cost and security.

Approach Cost visibility across clouds Security posture across clouds Joins cost and security on the same resource? Who maintains it Typical owner team
Native cloud consoles only (AWS Cost Explorer + Security Hub, Azure Cost Management + Defender for Cloud, GCP Billing + Security Command Center) Each cloud only Mostly per cloud; Defender for Cloud connects AWS and GCP, and Security Command Center Enterprise connects AWS and Azure No Each cloud provider Cloud team per provider
Microsoft Defender for Cloud with AWS and GCP connectors plus Azure Cost Management Azure only (Cost Management covers Microsoft Cloud costs) AWS, Azure, GCP No Microsoft Security and cloud ops
A FinOps platform alone (Vantage, CloudZero, Finout, or nOps) AWS, Azure, GCP; Vantage, CloudZero, and Finout also cover OCI None No FinOps vendor Finance or FinOps
A CNAPP alone (Wiz, Prisma Cloud, or Orca) None AWS, Azure, GCP, OCI No Security vendor Security
A FinOps platform plus a CNAPP AWS, Azure, GCP, and OCI with most FinOps platforms AWS, Azure, GCP, OCI Only if you join their outputs manually or via API Both vendors FinOps + Security
Observability platform with cost and security modules (e.g., Datadog Cloud Cost Management + Cloud Security) AWS, Azure, GCP, OCI AWS, Azure, GCP, OCI Partial, through shared tags across its separate cost and security products Observability vendor Platform engineering
FOCUS exports plus security findings in a data warehouse (BigQuery, Snowflake) with BI dashboards All four clouds export FOCUS All clouds with exportable findings Yes, if you build the join Your engineering team Data or platform engineering
Open-source tools (Steampipe for queries, Prowler for CSPM, OpenCost for Kubernetes cost) Partial, varies by plugin Partial, varies by tool Only if you build the join Community and your team Platform engineering

Datadog comes closest, covering both cost and security on all four clouds, but no commercial tool we found natively joins spend and security findings on the same resource. You either join two tools’ outputs or build the join in a warehouse.

What data does each cloud give you to work with?

Each major cloud provider offers billing exports, security findings, and asset inventory APIs, but the formats and capabilities differ.

Cloud Billing export (and FOCUS support) Security findings source Asset inventory source Tag or label limits to watch
AWS AWS Data Exports: CUR 2.0, plus separate FOCUS 1.0 and FOCUS 1.2 exports with AWS columns AWS Security Hub and Security Hub CSPM (aggregating GuardDuty, Inspector, Macie, and partner findings) AWS Config, AWS Resource Explorer 50 user tags per resource
Microsoft Azure Cost Management exports, including a cost and usage details (FOCUS) export Microsoft Defender for Cloud Azure Resource Graph 50 tags per resource
Google Cloud Cloud Billing export to BigQuery, including a FOCUS usage cost export (Preview); resource-level cost requires the detailed usage cost export Security Command Center Cloud Asset Inventory 64 labels per resource
Oracle Cloud Infrastructure (OCI) Cost reports, plus FOCUS-format CSV cost reports generated automatically Cloud Guard Search service 64 defined and 10 freeform tags per resource, 5 KB total tag data

All four clouds now export billing data in the FinOps Foundation FOCUS specification, which standardizes columns like BilledCost, EffectiveCost, ResourceId, and ServiceName, though column coverage can differ by provider. Tag limits matter because if you hit the ceiling, you may not be able to add the ownership and environment tags your pipeline requires. Plan your tag taxonomy early and keep it lean.

How do you join cost and security data on the same resources?

The same six steps work whether you buy tools and export their data or build everything from cloud-native exports.

Step What to do Output
1. One tagging and ownership standard enforced by policy Define required tags (owner, team, environment, service). Enforce them with AWS tag policies plus SCPs that deny creation when required tags are missing, Azure Policy deny rules, GCP custom Organization Policy constraints where the service supports them (plus checks in your infrastructure-as-code pipeline), and OCI tag defaults or required defined tags. Where a resource type cannot be blocked at creation, flag it in a daily compliance report. Every resource has a known owner and environment. Untagged resources are flagged or prevented.
2. Billing exports in FOCUS format Enable FOCUS billing exports for AWS (the Data Exports FOCUS table), Azure (the Cost Management FOCUS export), and Google Cloud (the FOCUS BigQuery export). For OCI, download the FOCUS-format cost reports. One billing table per cloud with consistent column names.
3. Security findings normalized Export findings from Security Hub, Defender for Cloud, Security Command Center, and Cloud Guard. Normalize to a common schema such as OCSF, or use a CNAPP that already aggregates across clouds. One findings table with severity, resource ID, cloud, and timestamp.
4. One asset inventory keyed by resource ID Pull resource inventories from AWS Config, Azure Resource Graph, Cloud Asset Inventory, and OCI Search. Normalize resource IDs and attach tags. A lookup table mapping resource ID to owner, team, environment, cloud, region, and resource type.
5. Join on resource ID and owner tag In your data warehouse or BI tool, join the billing table, findings table, and inventory table on a normalized resource ID (full ARN, lowercased Azure ID, Google Cloud full resource name, OCID). Not every billing line carries a resource ID: data transfer, support, tax, and commitment fees often do not, so allocate those by account, project, or owner tag instead. Use owner tag as a secondary join key for aggregation by team. A single table where each row has cost data, security findings, and ownership for one resource.
6. Dashboards and alerts per owner Build dashboards showing cost per finding severity, spend on resources with critical findings, untagged spend, and owner-level summaries. Set alerts for anomalies. Actionable views for FinOps, security, and engineering leads.

Once the join is in place, you can compute metrics that neither a FinOps tool nor a CNAPP can produce alone.

  • Spend on resources with critical findings. Sum the EffectiveCost of all resources that have at least one critical-severity finding open. This tells the CISO how much money is at risk and tells finance which cost centers carry the most security debt.
  • Cost of untagged resources. Sum the cost of resources missing the owner or team tag. This is both a governance metric and a cost-allocation gap.
  • Owner with the most findings per dollar spent. Divide the count of open findings by total spend for each owner tag. Weight it by severity and read it alongside absolute finding counts, since a small-budget team can show a high ratio from a handful of findings.
  • Mean time to remediate findings on high-spend resources. Track how long critical findings stay open on resources that cost more than a defined threshold. This metric aligns security response with business impact.
  • Unattached storage and idle compute cost. Join the asset inventory with billing data to find resources generating cost but no utilization metrics, and cross-reference with findings for exposed storage volumes.

What makes a single view hard to keep accurate?

Most of the work comes after launch, in keeping the pipeline accurate as accounts, teams, and resources change.

Tag drift. Engineers create resources without tags, rename teams without updating tags, or use inconsistent values like “prod” versus “production.” Enforcement at creation time helps, but you also need a weekly reconciliation job that flags tag violations.

Shared resources. A single database or load balancer may serve multiple teams. Allocating its cost and attributing its security findings to a single owner is inherently approximate. Document your allocation rules and make them visible to all stakeholders.

Kubernetes cost allocation. Containers share nodes, and node-level billing does not map cleanly to pod-level cost. OpenCost, a CNCF project, allocates node cost to pods and namespaces, but its output does not appear in cloud bills, so you join it to billing through the node’s resource ID. Our guide to cloud cost optimization tools compares the FinOps platforms that ingest it.

Currency and discount differences. AWS Reserved Instances and Savings Plans, Azure reservations and savings plans, Google Cloud committed use discounts, and OCI Universal Credits all reduce effective cost differently, and accounts billed in different currencies need conversion before you sum them. FOCUS normalizes some of this with BilledCost versus EffectiveCost, but discount attribution to specific resources can still be ambiguous.

Findings without owners. Security tools sometimes flag resources that have no tags, no known owner, and no clear cost attribution. These orphaned findings fall through the cracks unless you have a process to assign them to a default owner for triage.

Asset inventory lag. Cloud asset APIs update on different schedules. A resource can be created, generate cost, and be flagged by a security scanner before the inventory API reflects it. Design your pipeline to handle missing inventory records gracefully.

ID format mismatches. Billing and security tools often format the same resource differently, such as an EC2 instance ID in billing and a full ARN in Security Hub, or mixed-case Azure IDs. Normalize IDs in the inventory step, before the join.

How Control Plane fits

Control Plane is the unbreakable platform for running production workloads. You deploy a workload to a Control Plane Global Virtual Cloud (GVC), a named set of locations, and it runs in every location from one definition. Those locations can be on AWS, GCP, Azure, Oracle Cloud Infrastructure (OCI), Kubernetes clusters you already run, or your own machines, including on premises.

Every location of a GVC enforces the same controls. Workloads sit behind a deny-by-default internal firewall, talk to each other over mTLS, get permissions through policies attached to identities, and leave a tamper-proof audit trail of actions in the org. Control Plane’s Universal Cloud Identity (UCI) lets workloads reach AWS, GCP, and Azure services through one identity, with no long-lived cloud keys in your code. More on running one identity model across clouds.

On the cost side, Control Plane’s Cost & Usage view shows spend per org and per resource, with a projected total, across every cloud a GVC’s locations run on. A billing_viewer role gives finance read access to Cost & Usage and invoices, and a monthly spend alert emails every billing admin when spend reaches an amount you set. You pay for the CPU and memory a workload reserves, and Capacity AI, Control Plane’s automatic CPU and memory sizing, right-sizes reservations between a minimum and maximum based on historical usage.

Control Plane handles Day 2 operations for you, including autoscaling, built-in observability, and security. Control Plane is PCI DSS Level 1, SOC 2 Type II, HIPAA, and GDPR compliant. Related guides cover platforms that unify identity, billing, and deployment and hybrid cloud security and compliance risks.

Cost & Usage and the security controls above apply to every workload Control Plane runs, including databases deployed from its template catalog. Cloud services that run outside it, such as Amazon RDS, storage buckets, and managed queues, still go through the billing and findings exports in the pipeline above. Control Plane’s billing and usage APIs return the same numbers as the Cost & Usage view, grouped by GVC, workload, or location, so its spend loads into the same warehouse as one more pre-joined source.

FAQ

Answers to common questions about joining cost and security data across clouds.

How can I get a single view of cost and security across multiple clouds?

Enforce one tagging scheme, keep one asset inventory keyed by resource ID, and join billing and findings on that key, either by pairing a FinOps platform with a CNAPP or by building the join in a data warehouse.

What is FOCUS in FinOps?

FOCUS is the FinOps Open Cost and Usage Specification from the FinOps Foundation that standardizes cloud billing columns and semantics across providers. It makes it practical to query multi-cloud billing in one table.

Is a CNAPP the same as CSPM?

No. A CNAPP is a broader category that typically includes CSPM plus workload protection (CWPP), entitlement analysis (CIEM), and infrastructure-as-code scanning. CSPM focuses on misconfiguration and compliance checks.

Can one tool handle both FinOps and cloud security?

Partly. A few platforms, such as Datadog, cover both cost and security across AWS, Azure, Google Cloud, and OCI, but they show them as separate products. Attributing spend to open findings on the same resource is still a join you build or approximate with shared tags.

How do I allocate Kubernetes costs across clouds?

Use OpenCost for per-pod and per-namespace allocation and label pods and namespaces with the same ownership tags you use for cloud resources. Commercial FinOps tools can also ingest and reconcile Kubernetes allocation data, and native options exist: AWS split cost allocation data for EKS, GKE cost allocation in the Google Cloud billing export, and the AKS cost analysis add-on.

What is the best multi-cloud cost management tool?

There is no single best tool. Choose by fit: which clouds and Kubernetes platforms you use, whether the tool’s API supports exports for your join pipeline, and which team will own FinOps.

How does Control Plane show cost and security across clouds?

For workloads it runs, Control Plane puts spend and security controls in one place. Its Cost & Usage view breaks spend down per org and per resource with a projected total for the month. The same security controls apply in every location of a Control Plane Global Virtual Cloud (GVC), the set of locations a workload runs in, whether a location is on AWS, GCP, Azure, or Oracle Cloud Infrastructure (OCI). Those controls are a deny-by-default internal firewall, mTLS between workloads, identity-based access policies, and a tamper-proof audit trail. Capacity AI, Control Plane’s automatic CPU and memory sizing, right-sizes a workload’s CPU and memory reservation between a minimum and maximum based on historical usage, and that reservation is what CPU and memory are billed on.

Do I need a data warehouse to join cost and security data?

Not for small environments. If you have many accounts and need historical analysis, a warehouse like BigQuery or Snowflake is the scalable option for building and storing the joined dataset.

How often should I refresh the joined view?

Daily for billing is sufficient for most teams, since billing data typically lags 24 to 48 hours and can be restated until month close; security findings should refresh more frequently, ideally every few hours for critical issues. Asset inventory should refresh at least once a day.

To shrink the surface area you need to join, talk to the Control Plane team or sign up free.

Originally published on the Control Plane blog.