For AI agents: visit https://docs.umbrellacost.io/llms.txt for an index of all pages formatted in Markdown and endpoints in OpenAPI. Append .md to any documentation page URL to get its markdown version.
| Data type | Source(s) | Update frequency |
|---|
| Billing data | AWS Cost & Usage Report (CUR 2.0 and Legacy) delivered to S3 | Hourly |
| Price lists | AWS Price List Bulk API + Price List Query API | On-change (polled daily) |
| Resource inventory | Direct AWS API calls (EC2, RDS, S3, Lambda, etc.) | Near real-time; used for recommendations |
| Configurations | Direct AWS API calls (instance types, extended support, storage classes, networking, tags) | Near real-time (periodic polling) |
| K8s | CloudWatch agent | Hourly |
| K8s | Prometheus agent | Hourly |
| AI Latency / performance metrics | CloudWatch InvocationLatency per Bedrock model; Azure-hosted AI latency via native metrics | Per minute aggregated Hourly |
| Data type | Source(s) | Update frequency |
|---|
| Billing data | Azure Billing Export to Azure Storage Account via Cost Management | Daily |
| Price lists | Azure Retail Prices API | On-change (polled daily) |
| Resource inventory | Azure Resource Manager (ARM) API | Near real-time; used for recommendations |
| Configurations | Direct Azure API calls (VM SKUs, disk configs, networking, tags) | Near real-time |
| Service usage / consumption | Azure Billing Export (consumption meters); Azure Monitor and Azure Advisor (recommendations, K8s metrics) | Billing Export: daily; Monitor/Advisor: near real-time |
| K8s | Prometheus agent | Daily |
| AI Latency / performance metrics | Azure Monitor latency metrics per Azure OpenAI deployment | Per minute aggregated Daily |
| Data type | Source(s) | Update frequency |
|---|
| Billing data | BigQuery Billing Export (standard and detailed) | Daily (near real-time with detailed export) |
| Price lists | GCP Cloud Billing Catalog API | On-change (polled daily) |
| Resource inventory | GCP Cloud Asset Inventory / Resource Manager API | Near real-time; used for recommendations |
| Configurations | GCP Monitoring Client; direct API calls (Compute, GKE, Storage configs, labels) | Near real-time |
| Service usage / consumption | GCP Monitoring Client (K8s workload metrics, resource utilization, recommendations) | Near real-time |
| K8s + Recommendations | GCP Monitoring Client | Daily |
| AI Latency / performance metrics | Cloud Monitoring prediction latency per Vertex AI model | Per minute aggregated Hourly |
