Split multiple instances support

Currently, heartbeat backend does not support multiple instance which make it hard to scale up the instance when load increasing.

Multi-instance deployment is achieved through two methods: Docker Compose and k8s, and files can be shared between different instances.

File sharing between containers is achieved by introducing Volume in Docker Compose and PVC in k8s.

• docker-compose.yaml

version: "3.4"

services:
  backend:
    image: ghcr.io/au-heartbeat/heartbeat_backend:latest
    expose:
      - "4322"
    restart: always
    deploy:
      replicas: 3
    volumes:
      - file_volume:/app/output
  frontend:
    image: ghcr.io/au-heartbeat/heartbeat_frontend:latest
    container_name: frontend
    ports:
      - "4321:80"
    depends_on:
      - backend
    restart: always

volumes:
  file_volume:

Three backend containers are started here and the files are mounted to file_volume

Create a namespace if the user does not have one.

kubectl create namespace heartbeat

implement k8s yaml

• k8s-persistent-volume.yaml

apiVersion: v1
kind: PersistentVolume
metadata:
  name: pv
spec:
  capacity:
    storage: 200Mi
  accessModes:
    - ReadWriteMany
  storageClassName: heartbeat
  hostPath:
    path: /data

The spec.capacity.storage can be adjusted to the appropriate size.

• k8s-persistent-volume-claim.yaml

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: pvc
spec:
  accessModes:
    - ReadWriteMany
  resources:
    requests:
      storage: 200Mi
  storageClassName: heartbeat

The spec.resources.requests.storage can be adjusted to the appropriate size, but not larger than spec.capacity.storage in the k8s-persistent-volume.yaml.

• k8s-backend-deployment.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: backend
  labels:
    app: backend
spec:
  replicas: 3
  selector:
    matchLabels:
      app: backend
  template:
    metadata:
      labels:
        app: backend
    spec:
      initContainers:
        - name: backend-init
          image: busybox
          securityContext:
            runAsUser: 0
          volumeMounts:
            - name: volumes
              mountPath: /app/output
          command: ["sh", "-c", "chown -R 999:999 /app/output"]
      containers:
        - name: backend
          image: ghcr.io/au-heartbeat/heartbeat_backend:latest
          volumeMounts:
            - name: volumes
              mountPath: /app/output
      volumes:
        - name: volumes
          persistentVolumeClaim:
            claimName: pvc

Three backend pods are started here and the files are mounted to volumes

• k8s-backend-service.yaml

apiVersion: v1
kind: Service
metadata:
  name: backend
spec:
  selector:
    app: backend
  ports:
    - protocol: TCP
      port: 4322
      targetPort: 4322

• k8s-frontend-deployment.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: frontend
spec:
  selector:
    matchLabels:
      app: frontend
  template:
    metadata:
      labels:
        app: frontend
    spec:
      containers:
        - name: frontend
          image: ghcr.io/au-heartbeat/heartbeat_frontend:latest

• k8s-frontend-service.yaml

apiVersion: v1
kind: Service
metadata:
  name: frontend
spec:
  selector:
    app: frontend
  ports:
    - protocol: TCP
      port: 4321
      targetPort: 80
  type: LoadBalancer

When generating files, use ./app/output/csv instead of ./csv

• Mainly exists in classes

CSVFileGenerator.java
CSVFileNameEnum.java

• Check point

In the case of a single instance, check whether the three report files provided for download can be correctly generated and downloaded under the target file.

• AsyncExceptionHandler.java

  • Need to initialize a folder ./app/output/error.
  • Modify the put method to generate an intermediate file, the content of the file is the corresponding outlier, and then atomize the file name.
  • The get method reads the target file and returns null if the file does not exist, otherwise it returns an exception object.
  • When deleting a file fails, check if the file still exists. If it does not exist, the deletion is successful.

• AsyncReportRequestHandler.java

  • Need to initialize two folder ./app/output/report and ./app/output/metricsDataReady.
  • Modify the put method as AsyncExceptionHandler.
  • Modify the get method reads the target file and returns null if the file does not exist, otherwise it returns target object.
  • When deleting a file fails, check if the file still exists. If it does not exist, the deletion is successful.

• DeleteExpireCSVScheduler.java

  • Change file from ./csv/ to ./app/output/csv
  • When deleting a file fails, check if the file still exists. If it does not exist, the deletion is successful.

Put docker-compose.yaml in 7.1.1 Customize story point field in Jira in README