Multi-Cloud Cost Management in One Inventory
Three clouds bill three different ways. One inventory and one cost model is what makes them comparable.
Each cloud discovered its own way
ZopNight uses each provider’s native inventory service rather than a lowest-common-denominator scan:
- AWS: Resource Explorer, with per-service collectors for detail such as S3 bucket size and object counts and ECS capacity providers.
- GCP: Cloud Asset Inventory, plus Cloud Monitoring for Cloud Storage sizes and a dedicated collector for Artifact Registry.
- Azure: Resource Graph. A tenant-scoped connection finds every subscription it can reach and creates one account per subscription.
Databricks (on any of the three), Snowflake and Red Hat OpenShift on AWS connect as their own accounts. Parent and child resources are kept together, so a cluster, its node pools and their VMs read as one tree.
Connecting without long-lived secrets
Each cloud has a recommended keyless option: an IAM role assumed through workload identity federation on AWS, One-Click Connect through Google sign-in on GCP (ZopNight creates and manages a dedicated service account), and a service principal or workload identity federation on Azure. Static keys and service-account keys are supported when you need them.
Credentials can be rotated, or the method switched, without deleting the account. New credentials are checked against the provider first and must resolve to the same account, project or subscription, so an account can be re-authenticated but never silently repointed. Resources, schedules, recommendations and cost history stay attached.
Each account shows which permissions were granted or denied during discovery, by category, and ZopNight can alert you when a permission is newly denied.
One model on top
Once connected, the three clouds share the same features: the resource list with a single account filter grouped by provider, cost reports split by provider and by region within each provider, the same scheduling and override model, and one recommendations queue. Billed cost is read from each cloud’s own billing source, with Azure on amortized cost and GCP net of credits, so the figures line up with each invoice.
Connecting read-only gives you discovery, cost and recommendations; start/stop and remediation need read + write.
Related: multi-cloud cost strategy, AWS vs GCP vs Azure costs, cost reporting and Cloud IAM import.
Questions we get a lot.
Can I compare spend across clouds?
Yes. Cost reports break spend down by provider, account, region and type, and the Cost Flow view starts from provider by default.
What happens to my data if I rotate credentials?
Nothing is lost. Rotation and auth-method changes keep every resource, schedule, recommendation and cost record, and are refused if the new credential points at a different account.