Daily anomaly detection across 5 dimensions (org, cloud account, resource group, resource, team) using percent deviation and z-score. Higher severity wins. Root cause analysis identifies instance resizes, new resources, reservation expiry, and schedule failures per anomaly. ZopNight applies cost anomaly detection to non-production across AWS, GCP, and Azure, measured against real usage. See it by cloud or by service below, and the discipline in FinOps.
By cloud
By service
- Cost Anomaly Detection for EC2
- Cost Anomaly Detection for RDS
- Cost Anomaly Detection for EKS
- Cost Anomaly Detection for ECS
- Cost Anomaly Detection for Lambda
- Cost Anomaly Detection for ElastiCache
- Cost Anomaly Detection for Redshift
- Cost Anomaly Detection for SageMaker
- Cost Anomaly Detection for EMR
- Cost Anomaly Detection for ASG
- Cost Anomaly Detection for S3
- Cost Anomaly Detection for EBS
- Cost Anomaly Detection for CloudFront
- Cost Anomaly Detection for DynamoDB
- Cost Anomaly Detection for App Runner
- Cost Anomaly Detection for Compute Engine
- Cost Anomaly Detection for Cloud SQL
- Cost Anomaly Detection for GKE
- Cost Anomaly Detection for Cloud Run
- Cost Anomaly Detection for Memorystore
- Cost Anomaly Detection for BigQuery
- Cost Anomaly Detection for Cloud Functions
- Cost Anomaly Detection for Artifact Registry
- Cost Anomaly Detection for Azure VMs
- Cost Anomaly Detection for AKS
- Cost Anomaly Detection for Azure SQL
- Cost Anomaly Detection for Azure Functions
- Cost Anomaly Detection for VMSS
- Cost Anomaly Detection for Databricks
- Cost Anomaly Detection for ML Compute
Questions we get a lot.
If yours isn't here, email us and we'll answer directly.
Where should I start?
Start with the FinOps overview for the discipline that ties it together, then use the specific guide, cloud, or location page for the task you are on.
Is production at risk?
No. Production is excluded by default; ZopNight acts only on the non-production resources you choose.
What access does ZopNight need?
A read-only role to start. Discovery, cost reporting, and recommendations all run read-only, and you enable actions when you are ready.
Which clouds are supported?
AWS, GCP, and Azure from one platform, including Databricks across all three. Schedules, groups, overrides, and recommendations work the same way regardless of provider.