Metric types in Cloud Monitoring are classified into general groups, based on the type of service that collects the data. This document provides links to reference lists for each of these groups.
Metrics from Google Cloud services include:
- Google Cloud metrics, for Google Cloud services such as Compute Engine and BigQuery.
- Google Distributed Cloud metrics, for Google Distributed Cloud software only.
- Istio metrics, for Istio on Google Kubernetes Engine.
- Kubernetes metrics, for Google Kubernetes Engine (GKE).
Metrics for the legacy Cloud Monitoring and Cloud Logging agents, as well as the Ops Agent. With the exception of metrics from third-party applications, the Ops Agent and the legacy agents share many groups of metrics.
Ops Agent metrics, for VM instances running the Ops Agent.
Legacy Monitoring and Logging agent metrics, for VM instances running the legacy Monitoring and Logging agents.
Knative metrics, for Knative components.
External metrics, for open-source and third-party applications.
About the lists
The tables in the metric-type lists are regenerated frequently and time-stamped.
The information listed for each metric type comes from
the Monitoring API MetricDescriptor
object for each metric type. For more information about how metric
types are described, see
Metrics, time series, and resources.
Launch stages
Each metric type has a launch stage that indicates its maturity. If specified, the launch stage appears as a colored superscript after the metric type: GA, BETA, ALPHA, EARLY_ACCESS, or DEPRECATED. If no superscript appears, then the launch stage is unspecified. For more information, see Product launch stages.
Metric types in the Alpha or Early Access launch stages might not appear in
the public lists of metrics. To get information about those metric types,
explicitly retrieve the set of metric descriptors from a Google Cloud
project that has been given permission to use the restricted metric types.
If you have permission to use restricted metric types, you can retrieve
the metric descriptors by using the
metricDescriptors.list
method in the
Monitoring API. For more information, see
List metric descriptors.
Additional information: metadata
The description for each metric type includes information called metadata about the metric. The metadata includes the following:
Sample Period: For metrics that are written periodically, this is the time interval between consecutive data points, excluding data loss due to errors. The period, if available, appears at the end of the description text in a sentence of the form “Sampled every x seconds.”
Latency: Data points older than this value are guaranteed to be available to be read, excluding data loss due to errors. The delay does not include any time spent waiting until the next sampling period. The delay, if available, appears at the end of the description text in a sentence of the form “After sampling, data is not visible for up to y seconds.”
Resource hierarchy levels: Most metrics are collected only for projects, but some metrics are also collected at the organization and folder levels. If a level is not specified in the metadata, then the metric is written at the project level.
Pricing
In general, Cloud Monitoring system metrics are free, and metrics from external systems, agents, or applications are not. Billable metrics are billed by either the number of bytes or the number of samples ingested.
For more information about Cloud Monitoring pricing, see the following documents:
Getting started
- For an introduction to the concepts and terminology used in the Cloud Monitoring metric model, see Metrics, time series, and resources.
- To create your own metrics, see Create user-defined metrics with the API, User-defined agent metrics, and Logs-based metrics.
- To quickly see graphs of metric data, use the Metrics Explorer. For information about using this tool, see Create charts with Metrics Explorer.