When deploying an app that uses Pulse on Kubernetes, every time there's a new deployme

Option to ignore offline servers about pulse HOT 2 CLOSED

leventdev commented on July 30, 2024

Option to ignore offline servers

from pulse.

Comments (2)

driesvints commented on July 30, 2024

I think we'd appreciate a PR for this, thanks!

from pulse.

DellanX commented on July 30, 2024

Something to add to this, for future developers planning to add Pulse to their Kubernetes cluster, is to set the PULSE_SERVER_NAME to be the node name, which you can do by adding the following snippet to your container spec:

          env:
            - name: PULSE_SERVER_NAME
              valueFrom:
                fieldRef:
                  fieldPath: spec.nodeName

Furthermore, if you use a Daemon Set instead of a deployment, you can configure there to be exactly one pulse service per node.

This is my configuration:

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: {{ .Release.Name }}-pulse-check-default
  labels:
    tier: backend
    layer: pulse-check
spec:
  replicas: 1
  selector:
    matchLabels:
      tier: backend
      layer: pulse-check
  template:
    metadata:
      labels:
        tier: backend
        layer: pulse-check
    spec:
      imagePullSecrets:
        - name: {{ .Release.Name }}-registry-key
      containers:
        - name: pulse-check
          image: <docker-registry>:{{ .Values.version }}
          command:
            - php
          args:
            - artisan
            - pulse:check
          ports:
            - containerPort: 9000
          env:
            - name: PULSE_SERVER_NAME
              valueFrom:
                fieldRef:
                  fieldPath: spec.nodeName
          envFrom:
            - configMapRef:
                name: {{ .Release.Name }}
            - secretRef:
                name: {{ .Release.Name }}