Clouds to applications
One observability data layer

sum by (cluster, namespace) (
  rate(container_cpu_usage_seconds_total{container!=""}[5m])
)

Compare namespace pressure across clusters without changing the metric query language your dashboards use

sum by (cloud_region, faas_name) (
  rate(faas_invocations_total{status="error"}[5m])
)

Group serverless failures by region and function using standard PromQL

sum by (cloud_region, service_name) (
  rate(http_server_request_duration_seconds_count{http_response_status_code=~"5.."}[5m])
)

Aggregate application failures by service and Azure region

histogram_quantile(0.99,
  sum by (le, service_name) (
    rate(http_server_request_duration_seconds_bucket[5m])
  )
)

Calculate tail latency for instrumented services with standard PromQL

max by (db_system, server_address) (
  db_client_connections_usage{state="used"}
)

Compare connection usage across engines and instances using a common metric model

sum by (device, interface) (
  rate(network_interface_errors_total[5m])
)

Rank error rates across network devices and interfaces with PromQL

sum(rate(http_server_request_duration_seconds_count{http_response_status_code=~"5.."}[5m]))
/
sum(rate(http_server_request_duration_seconds_count[5m]))

Calculate a service error ratio using the same PromQL dashboards already expect

sum by (gen_ai_request_model) (
  rate(gen_ai_client_token_usage_total[5m])
)

Compare token volume by requested model using standard metric queries