Metrics and usage

What is collected, how cloud metrics are named and labelled, and how to use them in dashboards, alerts and Vera.

metricslabelsdashboardsalertsveracost

What is collected

VerOps collects numeric time-series metrics from the provider's monitoring service for the resources and namespaces you selected. Each datapoint is normalized to a consistent name and labelled with standard resource attributes.

Aspect AWS Azure
Source CloudWatch GetMetricData Azure Monitor metrics
Example name cloud.aws.ec2.cpu_utilization cloud.azure.compute_virtualmachines.percentage_cpu
Default statistic Average Average
Selected by Namespaces (+ optional metric list) Resource IDs (+ optional metric list)

Standard labels

Every cloud series carries these labels, so you can filter and group the same way regardless of provider:

  • cloud_provideraws or azure
  • cloud_account_id — AWS account id or Azure subscription id
  • cloud_region — region / location
  • cloud_service — e.g. ec2, rds, compute_virtualmachines
  • cloud_resource_id — the specific resource
  • plus the metric's own dimensions and the native metric name.

Using cloud metrics

  • Dashboards. Add a metric widget and search for cloud. to browse what's arriving.
  • Alerts. Build alert rules on cloud metrics exactly like any other metric.
  • Vera. Ask about a resource by name; cloud metrics are part of the context Vera reasons over.

Freshness and cost

Metrics are pulled on the integration's poll interval (default 5 minutes). Cloud providers also finalize recent datapoints with a short delay, so the newest minute may fill in on the next poll. A shorter interval means fresher data but more provider API calls on your cloud bill — tune it to taste.

Tip: The number of series each poll wrote is shown on the integration row. If it's higher than you expect, tighten the namespace / resource / metric selection to control cardinality and cost.