AppDynamics Operator simplifies the configuration and lifecycle management of the AppDynamics ClusterAgent and the AppDynamics Machine Agent on different Kubernetes distributions and OpenShift. The Operator encapsulates key operational knowledge on how to configure and upgrade the ClusterAgent and the Machine Agent. It knows, for example, which configuration changes are benign and do not require restart of the ClusterAgent, which minimizes unnecesary load on the cluster API server.

The Operator is implemented using OperatorSDK and uses Kubernetes API to maintain the desired state of the custom resources that represent the ClusterAgent and the Machine Agent. When the Operator is deployed, it creates custom resource definitions (CRDs) for 2 custom resources:

• clusteragent, which represents the ClusterAgent.
• infraviz, which represents the Machine Agent bundled with netviz and analytics.

This level of abstraction further simplifies the management of monitoring and imstrumentation and ensures granular security policy of the ClusterAgent and the Machine Agent.

Create namespace for the operator and the ClusterAgent

• Create namespace for AppDynamics components

  • Kubernetes kubectl create namespace appdynamics
  • OpenShift oc new-project appdynamics --description="AppDynamics Infrastructure"

• Create Secret cluster-agent-secret with the following key

  • The "controller-key" - the access key to the AppDynamics controller.

kubectl -n appdynamics create secret generic cluster-agent-secret \
--from-literal=controller-key="<controller-access-key>" \

• Update the image reference in the Operator deployment spec (deploy/cluster-agent-operator.yaml), if necessary.

The default is "docker.io/appdynamics/cluster-agent-operator:latest".

• Deploy the Operator

kubectl apply -f deploy/cluster-agent-operator.yaml

By default "docker.io/appdynamics/cluster-agent-operator:latest" is used.

AppDynamics images are also available from Red Hat Container Catalog.

To enable pulling, create a secret in the ClusterAgent namespace. In this example, namespace appdynamics is used and appdynamics-operator account is linked to the secret.

$ oc -n appdynamics create secret docker-registry redhat-connect
--docker-server=registry.connect.redhat.com
--docker-username=REDHAT_CONNECT_USERNAME
--docker-password=REDHAT_CONNECT_PASSWORD --docker-email=unused
$ oc -n appdynamics secrets link appdynamics-operator redhat-connect
--for=pull

The AppDynamics Cluster Agent is a new, lightweight agent written in Golang providing APM monitoring for Kubernetes. The Cluster Agent is designed to help you understand how Kubernetes infrastructure affects your applications and business performance. With the AppDynamics Cluster Agent, you can collect metadata, metrics, and events about a Kubernetes cluster. You can query for the components in the cluster, and report metrics for those components. The Cluster Agent works across cloud platforms such as Kubernetes on AWS EKS, AKS on Azure, PKS on Pivotal and many others. The clusteragent is the custom resource that the Operator works with to deploy an instance of the ClusterAgent. When a clusteragent spec is provided, the Operator will create a single replica deployment and the necessary additional resources (a configMap and a service) to support the ClusterAgent.

Here is an example of a minimalistic spec of the ClusterAgent custom resource:

apiVersion: appdynamics.com/v1alpha1
kind: Clusteragent
metadata:
  name: k8s-cluster-agent
  namespace: appdynamics
spec:
  appName: "<app-name>"
  controllerUrl: "<protocol>://<appdynamics-controller-host>:<port>"
  account: appdynamics-cluster-agent
  image: "<your-docker-registry>/appdynamics/cluster-agent:tag"

Update the provided spec with the AppDynamics account information and deploy:

kubectl apply -f deploy/cluster-agent.yaml

Example resource limits:

   resources:
    limits:
      cpu: 300m
      memory: "200M"
    requests:
      cpu: 200m
      memory: "100M"

Here is an example of the entire spec of the ClusterAgent custom resource:

    apiVersion: appdynamics.com/v1alpha1
    kind: Clusteragent
    metadata:
      name: k8s-cluster-agent
      namespace: appdynamics
    spec:
      appName: "<app-name>"
      controllerUrl: "<protocol>://<appdynamics-controller-host>:<port>"
      account: "<account-name>"
      # docker image info
      image: "<your-docker-registry>/appdynamics/cluster-agent:tag"
      nsToMonitor:
        - "default"
      eventUploadInterval: 10
      containerRegistrationInterval: 120
      httpClientTimeout: 30
      customSSLSecret: "<secret-name>"
      proxyUrl: "<protocol>://<domain>:<port>"
      proxyUser: "<proxy-user>"
      metricsSyncInterval: 30
      clusterMetricsSyncInterval: 60
      metadataSyncInterval: 60
      containerBatchSize: 25
      podBatchSize: 30
      metricUploadRetryCount: 3
      metricUploadRetryIntervalMilliSeconds: 5
      podFilter:
        blacklistedLabels:
          - label1: value1
        whitelistedLabels:
          - label1: value1
          - label2: value2
        whitelistedNames:
          - name1
        blacklistedNames:
          - name1
          - name2
      logLevel: "INFO"
      logFileSizeMb: 5
      logFileBackups: 3
      stdoutLogging: "true"

Appdynamics operator can be used to enable server and network visibility with AppDynamics Machine agent. The operator works with custom resource infraviz to deploy the AppDynamics Machine Agent daemon set.

Here is an example of a minimalistic infraviz spec with the required parameters:

apiVersion: appdynamics.com/v1alpha1
kind: InfraViz
metadata:
  name: appd-infraviz
  namespace: appdynamics
spec:
  controllerUrl: "https://appd-controller.com"
  image: "docker.io/appdynamics/machine-agent-analytics:latest"
  account: "<your-account-name>"
  globalAccount: "<your-global-account-name"

The controller URL must be in the following format: <protocol>://<controller-domain>:<port>

Use the provided specs for Kubernetes and OpenShift to deploy the Machine Agent. Make sure to update the spec with the AppDynamics account information prior to deployment.

kubectl apply -f deploy/infraviz.yaml

On OpenShift:

oc apply -f deploy/infraviz-openshift.yaml

Example resource limits:

   resources:
    limits:
      cpu: 600m
      memory: "1G"
    requests:
      cpu: 300m
      memory: "800M"