To run a custom check, you can configure the DatadogAgent resource to provide custom checks (checks.d) and their corresponding configuration files (conf.d) at initialization time. You must configure a ConfigMap resource for each check script file and its configuration file.

This page explains how to set up a custom check, hello, that submits a hello.world metric to Datadog.

To learn more about checks in the Datadog ecosystem, see Introduction to Integrations. To configure a Datadog Integration, see Kubernetes and Integrations

Create the check files

Each check needs a configuration file (hello.yaml) and a script file (hello.py).

1. Create hello.yaml with the following content:

   init_config:
   
   instances: [{}]
   

2. Create hello.py with the following content:

   from datadog_checks.base import AgentCheck
   
   __version__ = "1.0.0"
   class HelloCheck(AgentCheck):
       def check(self, instance):
           self.gauge('hello.world', 1, tags=['env:dev'])
   

Create the check ConfigMaps

After you create the hello check files, create the associated ConfigMaps:

1. Create the ConfigMap for the custom check YAML configuration file hello.yaml:

   $ kubectl create configmap -n $DD_NAMESPACE confd-config --from-file=hello.yaml
   configmap/confd-config created
   

2. Verify that the ConfigMap has been correctly created:

   $ kubectl get configmap -n $DD_NAMESPACE confd-config -o yaml
   apiVersion: v1
   data:
     hello.yaml: |
       init_config:
   
       instances: [{}]
   kind: ConfigMap
   metadata:
     name: confd-config
     namespace: datadog
   

3. Create the ConfigMap for the custom check Python file hello.py:

   $ kubectl create configmap -n $DD_NAMESPACE checksd-config --from-file=hello.py
   configmap/checksd-config created
   

4. Verify that the ConfigMap has been correctly created:

   $ kubectl get configmap -n $DD_NAMESPACE checksd-config -o yaml
   apiVersion: v1
   data:
     hello.py: |
      from datadog_checks.base import AgentCheck
   
      __version__ = "1.0.0"
      class HelloCheck(AgentCheck):
        def check(self, instance):
          self.gauge('hello.world', 1, tags=['env:dev'])
    kind: ConfigMap
    metadata:
      name: checksd-config
      namespace: datadog
   

Configure the Datadog Agent

After you create your ConfigMaps, create a DatadogAgent resource to use them:

apiVersion: datadoghq.com/v2alpha1
kind: DatadogAgent
metadata:
  name: datadog
spec:
  global:
    credentials:
      apiKey: "<DATADOG_API_KEY>"
      appKey: "<DATADOG_APP_KEY>"
  override:
    nodeAgent:
      extraConfd:
        configMap:
          name: confd-config
      extraChecksd:
        configMap:
          name: checksd-config

Note: Any ConfigMaps you create need to be in the same DD_NAMESPACE as the DatadogAgent resource.

This deploys the Datadog Agent with your custom check.

ConfigMaps for multiple checks

You can populate ConfigMaps with the content of multiple checks or their respective configuration files.

Populating all check script files

$ kubectl create cm -n $DD_NAMESPACE checksd-config $(find ./checks.d -name "*.py" | xargs -I'{}' echo -n '--from-file={} ')
configmap/checksd-config created

Populating all check configuration files

$ kubectl create cm -n $DD_NAMESPACE confd-config $(find ./conf.d -name "*.yaml" | xargs -I'{}' echo -n '--from-file={} ')
configmap/confd-config created

Provide additional volumes

You can mount additional user-configured volumes in either the node or Cluster Agent containers by setting the volumes and volumeMounts properties.

Example: Using a volume to mount a secret

apiVersion: datadoghq.com/v2alpha1
kind: DatadogAgent
metadata:
  name: datadog
spec:
  global:
    credentials:
      apiKey: "<DATADOG_API_KEY>"
      appKey: "<DATADOG_APP_KEY>"
  override:
    nodeAgent:
      image:
        name: "gcr.io/datadoghq/agent:latest"
      volumes:
        - name: secrets
          secret:
            secretName: secrets
      containers:
        agent:
          volumeMounts:
            - name: secrets
              mountPath: /etc/secrets
              readOnly: true