How to deploy Site24x7 Python APM agent in Kubernetes using Init Containers

Adding the APM Insight Python agent in Kubernetes via init containers

Step 1: Create Kubernetes secret for Site24x7 license key

Create a secret for the Site24x7 license key in your application namespace.
Notes
You can obtain the license key from your Site24x7 account by navigating to Admin > Developer > Device Key.
Example
  1. kubectl create secret generic app-secret --from-literal=S247_LICENSE_KEY='your_s247_license_key' -n pythonapp
Replace app-secret, your_s247_license_key, and pythonapp (namespace) with the appropriate values.

Step 2: Define empty volume for APM agent files

Create an empty volume to copy the agent files during the initContainers process.
Example
  1. volumes:
  2.      - name: s247agent

Step 3: Add initContainer to copy APM agent files

Include the following initContainers command in your Helm chart or deployment YAML file.
  1. initContainers:
  2.   - name: agent-copy
  3.     image: site24x7/apminsight-pythonagent:latest
  4.     imagePullPolicy: IfNotPresent
  5.     command: ['cp', '-r', '/opt/site24x7/.', '/home/apm']
  6.     volumeMounts:
  7.       - name: s247agent
  8.         mountPath: /home/apm

Step 4: Mount agent volume into application container

Mount the volume created in Step 2 into your application container.
Example
  1. containers:
  2.   - name: pythonapp
  3.     image: pythonapp:latest
  4.     imagePullPolicy: IfNotPresent
  5.     ports:
  6.       - containerPort: 8080
  7.     volumeMounts:
  8.       - name: s247agent
  9.         mountPath: /home/apm

Step 5: Configure environment variables for APM

Add the following environment variables to the application containers.

Environment variable 1:
Name: S247_LICENSE_KEY 
Value: Enter the s247licensekey from the secret added in Step 1.

Environment variable 2:
Name: APM_APP_NAME 
Value: Configure the name of the Python application that will be displayed in the Site24x7 client.

Environment variable 3 (Optional):
Name: APP_ENV_PATH
Value: If you are using a virtual environment, specify the path of the environment bin directory.

Environment variable 4 (Optional):
Name: APM_LOGS_DIR
Value: Specify the path where your application log files are stored for log collection.
Notes
In this step, we configure the application (Python process) startup command as an argument so that the agent will send data to the specified monitor name.

Step 6: Set application startup command

Configure the application startup command as an environment variable.

To start your application together with the Python agent, you must provide your application’s startup command as an environment variable named APP_RUN_COMMAND.

This command should be the same as the CMD or ENTRYPOINT instruction defined at the end of your Dockerfile.

For example, if your original Dockerfile contains:
  1. CMD ["gunicorn", "--bind", "127.0.0.1:8080" "-w", "2", "pythonapp_main:app"]
Set the environment variable as follows:
  1. env:
  2.   - name: APP_RUN_COMMAND
  3.     value: "gunicorn --bind 127.0.0.1:8080 -w 2 pythonapp_main:app"
Example
Here is an example of a Python application running in a Gunicorn environment on port 8080 with two worker processes.
  1. env:
  2.   - name: APP_RUN_COMMAND
  3.     value: gunicorn --bind 127.0.0.1:8080 -w 2 pythonapp_main:app

  5.   - name: S247_LICENSE_KEY
  6.     valueFrom:
  7.       secretKeyRef:
  8.         name: pythonapp-secrets
  9.         key: S247_LICENSE_KEY

Step 7: Update container startup to use APM agent script

Change your application container startup command to the Python agent startup script.
Notes
The script agent_start.sh will start your Python application with the APM Python agent using the command provided in APP_RUN_COMMAND.
  1. containers:
  2.   - name: pythonapp
  3.     image: pythonapp:latest
  4.     imagePullPolicy: IfNotPresent
  5.     command: ["/bin/sh", "-c", "/home/apm/agent_start.sh"]
  6.     ports:
  7.       - containerPort: 8080
  8.     volumeMounts:
  9.       - name: s247agent
  10.         mountPath: /home/apm
Example YAML deployment file for your reference
  1. apiVersion: v1
  2. kind: Namespace
  3. metadata:
  4.   name: pythonapp
  5. ---
  6. apiVersion: v1
  7. kind: Secret
  8. metadata:
  9.   name: pythonapp-secrets
  10.   namespace: pythonapp
  11. type: Opaque
  12. data:
  13.   S247_LICENSE_KEY: dXNCGfNGPfNTMyZjVjNzAwOTE5YTM4ODQyMDQzOGVjZjAwNGE4NA==
  14. ---
  15. apiVersion: apps/v1
  16. kind: Deployment
  17. metadata:
  18.   name: pythonapp-deployment
  19.   namespace: pythonapp
  20.   labels:
  21.     app: flask
  22. spec:
  23.   replicas: 1
  24.   selector:
  25.     matchLabels:
  26.       app: flask
  27.   template:
  28.     metadata:
  29.       labels:
  30.         app: flask
  31.     spec:
  32.       initContainers:
  33.         - name: agent-copy-init
  34.           image: site24x7/apminsight-pythonagent:latest
  35.           imagePullPolicy: Always
  36.           command: ["cp", "-r", "/opt/site24x7/.", "/home/apm"]
  37.           volumeMounts:
  38.             - name: s247agent
  39.               mountPath: /home/apm

  41.       containers:
  42.         - name: pythonapp
  43.           image: pythonapp:version1
  44.           command: ["/bin/sh", "-c", "/home/apm/agent_start.sh"]
  45.           imagePullPolicy: Always
  46.           ports:
  47.             - containerPort: 5000
  48.           env:
  49.             - name: APP_RUN_COMMAND
  50.               value: "gunicorn --bind 127.0.0.1:8080 -w 2 pythonapp_main:app"
  51.             - name: APM_APP_NAME
  52.               value: "Python-application"
  53.             - name: S247_LICENSE_KEY
  54.               valueFrom:
  55.                 secretKeyRef:
  56.                   name: pythonapp-secrets
  57.                   key: S247_LICENSE_KEY
  58.           volumeMounts:
  59.             - name: s247agent
  60.               mountPath: /home/apm

  62.       volumes:
  63.         - name: s247agent
  64.           emptyDir: {}

  66.       restartPolicy: Always
  67. ---
  68. apiVersion: v1
  69. kind: Service
  70. metadata:
  71.   name: pythonapp-deployment
  72.   namespace: pythonapp
  73. spec:
  74.   type: NodePort
  75.   selector:
  76.     app: flask
  77.   ports:
  78.     - protocol: TCP
  79.       port: 5000
  80.       targetPort: 5000
  81.       nodePort: 30200