Overview

Collect your exposed Prometheus and OpenMetrics metrics from your application running inside Kubernetes by using the Datadog Agent and the OpenMetrics or Prometheus integrations. By default, all metrics retrieved by the generic Prometheus check are considered custom metrics.

Starting with version 6.5.0, the Agent includes OpenMetrics and Prometheus checks capable of scraping Prometheus endpoints. Datadog recommends using the OpenMetrics check since it is more efficient and fully supports the Prometheus text format. For more advanced usage of the OpenMetricsCheck interface, including writing a custom check, see the Developer Tools section. Use the Prometheus check only when the metrics endpoint does not support a text format.

This page explains the basic usage of these checks, which enable you to scrape custom metrics from Prometheus endpoints. For an explanation of how Prometheus and OpenMetrics metrics map to Datadog metrics, see the Mapping Prometheus Metrics to Datadog Metrics guide.

Setup

Installation

Deploy the Datadog Agent in your Kubernetes cluster. OpenMetrics and Prometheus checks are included in the Datadog Agent package, so you don’t need to install anything else on your containers or hosts.

Configuration

Configure your OpenMetrics or Prometheus check using Autodiscovery, by applying the following annotations to your pod exposing the OpenMetrics/Prometheus metrics:

# (...)
metadata:
  #(...)
  annotations:
    ad.datadoghq.com/<CONTAINER_IDENTIFIER>.checks: |
      {
        "openmetrics": {
          "init_config": {},
          "instances": [
            {
              "openmetrics_endpoint": "http://%%host%%:%%port%%/<PROMETHEUS_ENDPOINT> ",
              "namespace": "<METRICS_NAMESPACE_PREFIX_FOR_DATADOG>",
              "metrics": [{"<METRIC_TO_FETCH>":"<NEW_METRIC_NAME>"}]
            }
          ]
        }
      }      

spec:
  containers:
    - name: '<CONTAINER_IDENTIFIER>'

# (...)
metadata:
  #(...)
  annotations:
    ad.datadoghq.com/<CONTAINER_IDENTIFIER>.check_names: |
      ["openmetrics"]      
    ad.datadoghq.com/<CONTAINER_IDENTIFIER>.init_configs: |
      [{}]      
    ad.datadoghq.com/<CONTAINER_IDENTIFIER>.instances: |
      [
        {
          "openmetrics_endpoint": "http://%%host%%:%%port%%/<PROMETHEUS_ENDPOINT> ",
          "namespace": "<METRICS_NAMESPACE_PREFIX_FOR_DATADOG>",
          "metrics": [{"<METRIC_TO_FETCH>":"<NEW_METRIC_NAME>"}]
        }
      ]      
spec:
  containers:
    - name: '<CONTAINER_IDENTIFIER>'

With the following configuration placeholder values:

The metrics configuration is a list of metrics to retrieve as custom metrics. Include each metric to fetch and the desired metric name in Datadog as key value pairs, for example, {"<METRIC_TO_FETCH>":"<NEW_METRIC_NAME>"}. To prevent excess custom metrics charges, Datadog recommends limiting the scope to only include metrics that you need. You can alternatively provide a list of metric names strings, interpreted as regular expressions, to bring the desired metrics with their current names. If you want all metrics, then use ".*" rather than "*".

Note: Regular expressions can potentially send a lot of custom metrics.

For a full list of available parameters for instances, including namespace and metrics, see the sample configuration openmetrics.d/conf.yaml.

Getting started

Simple metric collection

1. Launch the Datadog Agent.

2. Use the Prometheus prometheus.yaml to launch an example Prometheus Deployment with the Autodiscovery configuration on the pod:

   Note: AD Annotations v2 was introduced in Datadog Agent version 7.36 to simplify integration configuration. For previous versions of the Datadog Agent, use AD Annotations v1.

     # (...)
    spec:
      template:
        metadata:
          annotations:
            ad.datadoghq.com/prometheus-example.checks: |
              {
                "openmetrics": {
                  "instances": [
                    {
                      "openmetrics_endpoint": "http://%%host%%:%%port%%/metrics",
                      "namespace": "documentation_example_kubernetes",
                      "metrics": [
                          {"promhttp_metric_handler_requests": "handler.requests"},
                          {"promhttp_metric_handler_requests_in_flight": "handler.requests.in_flight"},
                          "go_memory.*"
                        ]
                    }
                  ]
                }
              }           
        spec:
          containers:
          - name: prometheus-example
          # (...)
   

     # (...)
    spec:
      template:
        metadata:
          annotations:
            ad.datadoghq.com/prometheus-example.check_names: |
              ["openmetrics"]           
            ad.datadoghq.com/prometheus-example.init_configs: |
              [{}]           
            ad.datadoghq.com/prometheus-example.instances: |
              [
                {
                  "openmetrics_endpoint": "http://%%host%%:%%port%%/metrics",
                  "namespace": "documentation_example_kubernetes",
                  "metrics": [
                    {"promhttp_metric_handler_requests": "handler.requests"},
                    {"promhttp_metric_handler_requests_in_flight": "handler.requests.in_flight"},
                    "go_memory.*"
                  ]
                }
              ]           
        spec:
          containers:
          - name: prometheus-example
          # (...)
   

   Command to create the Prometheus Deployment:

   kubectl create -f prometheus.yaml
   

3. Go into your Metric summary page to see the metrics collected from this example pod. This configuration will collect the metric promhttp_metric_handler_requests, promhttp_metric_handler_requests_in_flight, and all exposed metrics starting with go_memory.

Metric collection with Prometheus annotations

With Prometheus Autodiscovery, the Datadog Agent is able to detect native Prometheus annotations (for example: prometheus.io/scrape, prometheus.io/path, prometheus.io/port) and schedule OpenMetrics checks automatically to collect Prometheus metrics in Kubernetes.

Requirements

• Datadog Agent v7.27+ or v6.27+ (for Pod checks)
• Datadog Cluster Agent v1.11+ (for service and endpoint checks)

Configuration

It’s recommended to first check which pods and services have the prometheus.io/scrape=true annotation before enabling this feature. This can be done with the following commands:

kubectl get pods -o=jsonpath='{.items[?(@.metadata.annotations.prometheus\.io/scrape=="true")].metadata.name}' --all-namespaces

kubectl get services -o=jsonpath='{.items[?(@.metadata.annotations.prometheus\.io/scrape=="true")].metadata.name}' --all-namespaces

Once the Prometheus scrape feature is enabled, the Datadog Agent collects custom metrics from these resources. If you do not want to collect the custom metrics from these resources, you can remove this annotation or update the Autodiscovery rules as described in the advanced configuration section.

Note: Enabling this feature without advanced configuration can cause a significant increase in custom metrics, which can lead to billing implications. See the advanced configuration section to learn how to only collect metrics from a subset of containers/pods/services.

Basic configuration

datadog:
  # (...)
  prometheusScrape:
    enabled: true
    serviceEndpoints: true
  # (...)

- name: DD_PROMETHEUS_SCRAPE_ENABLED
  value: "true"
- name: DD_PROMETHEUS_SCRAPE_VERSION
  value: "2"

- name: DD_PROMETHEUS_SCRAPE_ENABLED
  value: "true"
- name: DD_PROMETHEUS_SCRAPE_SERVICE_ENDPOINTS
  value: "true"

This instructs the Datadog Agent to detect the pods that have native Prometheus annotations and generate corresponding OpenMetrics checks.

It also instructs the Datadog Cluster Agent (if enabled) to detect the services that have native Prometheus annotations and generate corresponding OpenMetrics checks.

• prometheus.io/scrape=true: Required.
• prometheus.io/path: Optional, defaults to /metrics.
• prometheus.io/port: Optional, default is %%port%%, a template variable that is replaced by the container/service port.

This configuration generates a check that collects all metrics exposed using the default configuration of the OpenMetrics integration.

Advanced configuration

kubernetes_annotations:
  include:
     prometheus.io/scrape: "true"
  exclude:
     prometheus.io/scrape: "false"

datadog:
  # (...)
  prometheusScrape:
    enabled: true
    serviceEndpoints: true
    additionalConfigs:
      -
        configurations:
        - timeout: 5
          send_distribution_buckets: true
        autodiscovery:
          kubernetes_container_names:
            - my-app
          kubernetes_annotations:
            include:
              app: my-app

- name: DD_PROMETHEUS_SCRAPE_CHECKS
  value: "[{\"autodiscovery\":{\"kubernetes_annotations\":{\"exclude\":{\"prometheus.io/scrape\":\"false\"},\"include\":{\"prometheus.io/scrape\":\"true\"}}}}]"

- name: DD_PROMETHEUS_SCRAPE_ENABLED
  value: "true"
- name: DD_PROMETHEUS_SCRAPE_CHECKS
  value: "[{\"autodiscovery\":{\"kubernetes_annotations\":{\"include\":{\"app\":\"my-app\"}},\"kubernetes_container_names\":[\"my-app\"]},\"configurations\":[{\"send_distribution_buckets\":true,\"timeout\":5}]}]"
- name: DD_PROMETHEUS_SCRAPE_VERSION
  value: "2"

From custom to official integration

By default, all metrics retrieved by the generic Prometheus check are considered custom metrics. If you are monitoring off-the-shelf software and think it deserves an official integration, don’t hesitate to contribute!

Official integrations have their own dedicated directories. There’s a default instance mechanism in the generic check to hardcode the default configuration and metrics metadata. For example, reference the kube-proxy integration.

Further Reading

Additional helpful documentation, links, and articles:

Collect your application logsDOCUMENTATION Collect your application tracesDOCUMENTATION Automatically collect your applications' metrics and logsDOCUMENTATION Limit data collection to a subset of containers onlyDOCUMENTATION Assign tags to all data emitted by a containerDOCUMENTATION Mapping Prometheus Metrics to Datadog MetricsDOCUMENTATION