AWS App Mesh

Use the instructions below to enable metric collection for the AWS App Mesh proxy sidecar, called Envoy. Users can choose to add sidecars in one of three modes: deploying, patching the deployment later, or using the AWS App Mesh injector controller. All modes are supported by the following steps.

Metric collection

Prerequisite: Deploy Datadog Agents as a DaemonSet in your Kubernetes cluster using the EKS integration documentation.

1. Create a ConfigMap in your cluster to automatically discover App Mesh’s Envoy side cars that are added to each pod:

     apiVersion: v1
     kind: ConfigMap
     metadata:
     name: datadog-config
     data:
     envoy: |-
       ad_identifiers:
       - aws-appmesh-envoy
       init_config:
       instances:
       - stats_url: http://%%host%%:9901/stats
         tags:
           - <TAG_KEY>:<TAG_VALUE>  # Example - cluster:eks-appmesh    
   

2. Update the volumeMounts object in your Datadog Agent’s DaemonSet YAML file:

         volumeMounts:
          - name: datadog-config
            mountPath: /conf.d
   

3. Update the volumes object in your Datadog Agent’s DaemonSet YAML file:

        volumes:
         - name: datadog-config
           configMap:
             name: datadog-config
             items:
             - key: envoy
               path: envoy.yaml
   

Log collection

Trace collection

Select the namespace to deploy the datadog-agent and service, for example: monitoring. Use this in the option to deploy the appmesh-injector with:

```shell
  helm upgrade -i appmesh-controller eks/appmesh-controller \
  --namespace appmesh-system \
  --set sidecar.logLevel=debug \
  --set tracing.enabled=true \
  --set tracing.provider=datadog \
  --set tracing.address=ref:status.hostIP \
  --set tracing.port=8126
```

Alternatively, the appmesh injector can be deployed by following the App Mesh with EKS documentation using the option enable-datadog-tracing=true or environment variable ENABLE_DATADOG_TRACING=true.

Metric collection

Prerequisite: Add Datadog Agents to each of your Fargate task definitions with App Mesh enabled, such as an Envoy sidecar injected, using the ECS Fargate integration documentation.

1. To forward metrics from an ECS task with App Mesh to Datadog, follow the AWS App Mesh proposed model. In this approach, an external service is created as a virtual node to route outbound traffic.

   • In the App Mesh console, browse to your mesh.

   • Create a virtual node with the following details:

     Parameter	Value
     Virtual node name	datadog
     Service discovery type	DNS
     Hostname	app.datadoghq.com or app.datadoghq.eu
     Listener port	443
     Listener protocol	tcp

   • Create a virtual service with the following details:

     Parameter	Value
     Virtual service name	app.datadoghq.com or app.datadoghq.eu
     Provider	Virtual node
     Virtual node	datadog

2. Update all task definitions containing the Envoy sidecar and Datadog Agent with the following Docker labels. See Integration Setup for ECS Fargate for details.

       "dockerLabels": {
             com.datadoghq.ad.instances : [{"stats_url": "http://%%host%%:9901/stats"}]
             com.datadoghq.ad.check_names : ["envoy"]
             com.datadoghq.ad.init_configs : [{}]
           },
   

Log collection

Trace collection

1. Enable trace collection with the instructions in the ECS Fargate integration documentation.

Set the AWS App Mesh parameters ENABLE_ENVOY_DATADOG_TRACING and DATADOG_TRACER_PORT as environment variables in the ECS Fargate task definition. Learn more in the AWS App Mesh documentation.

Metric collection

Prerequisite: Add Datadog Agents to each of your ECS EC2 task definitions with App Mesh enabled, such as an Envoy sidecar injected, using the ECS integration documentation.

1. To forward metrics from an ECS task with App Mesh to Datadog, follow the AWS App Mesh proposed model. In this approach, an external service is created as a virtual node to route outbound traffic.

   • In the App Mesh console, browse to your mesh.

   • Create a virtual node with the following details:

     Parameter	Value
     Virtual node name	datadog
     Service discovery type	DNS
     Hostname	app.datadoghq.com or app.datadoghq.eu
     Listener port	443
     Listener protocol	tcp

   • Create a virtual service with the following details:

     Parameter	Value
     Virtual service name	app.datadoghq.com or app.datadoghq.eu
     Provider	Virtual node
     Virtual node	datadog

2. Update all task definitions containing the Envoy sidecar and Datadog Agent with the following Docker labels. See Integration Setup for ECS Fargate for details.

       "dockerLabels": {
             com.datadoghq.ad.instances : [{"stats_url": "http://%%host%%:9901/stats"}]
             com.datadoghq.ad.check_names : ["envoy"]
             com.datadoghq.ad.init_configs : [{}]
           },
   

Log collection

Trace collection

1. Enable trace collection with the instructions in the ECS integration documentation.

2. Set the AWS App Mesh parameters ENABLE_ENVOY_DATADOG_TRACING and DATADOG_TRACER_PORT as environment variables in the ECS task definition. Learn more in the AWS App Mesh documentation.