cannaiq/k8s/scraper-worker.yaml

# Task Worker Deployment
#
# Simple Deployment that runs task-worker.js to process tasks from worker_tasks queue.
# Workers pull tasks using DB-level locking (FOR UPDATE SKIP LOCKED).
#
# The worker will wait up to 60 minutes for active proxies to be added before failing.
# This allows deployment to succeed even if proxies aren't configured yet.
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: scraper-worker
  namespace: dispensary-scraper
spec:
  replicas: 25
  selector:
    matchLabels:
      app: scraper-worker
  template:
    metadata:
      labels:
        app: scraper-worker
    spec:
      imagePullSecrets:
        - name: regcred
      containers:
        - name: worker
          image: code.cannabrands.app/creationshop/dispensary-scraper:latest
          command: ["node"]
          args: ["dist/tasks/task-worker.js"]
          envFrom:
            - configMapRef:
                name: scraper-config
            - secretRef:
                name: scraper-secrets
          env:
            - name: WORKER_MODE
              value: "true"
            - name: POD_NAME
              valueFrom:
                fieldRef:
                  fieldPath: metadata.name
          resources:
            requests:
              memory: "256Mi"
              cpu: "100m"
            limits:
              memory: "512Mi"
              cpu: "500m"
          livenessProbe:
            exec:
              command:
                - /bin/sh
                - -c
                - "pgrep -f 'task-worker' > /dev/null"
            initialDelaySeconds: 60
            periodSeconds: 30
            failureThreshold: 3
      terminationGracePeriodSeconds: 60
---
# =============================================================================
# ALTERNATIVE: StatefulSet with multiple workers per pod (not currently used)
# =============================================================================
# Task Worker Pods (StatefulSet)
# Each pod runs 5 role-agnostic workers that pull tasks from worker_tasks queue.
#
# Architecture:
# - Pods are named from a predefined list (Aethelgard, Xylos, etc.)
# - Each pod spawns 5 worker processes
# - Workers register with API and show their pod name
# - HPA scales pods 5-15 based on pending task count
# - Workers use DB-level locking (FOR UPDATE SKIP LOCKED) to prevent conflicts
#
# Pod Names (up to 25):
# Aethelgard, Xylos, Kryll, Coriolis, Dimidium, Veridia, Zetani, Talos IV,
# Onyx, Celestia, Gormand, Betha, Ragnar, Syphon, Axiom, Nadir, Terra Nova,
# Acheron, Nexus, Vespera, Helios Prime, Oasis, Mordina, Cygnus, Umbra
---
apiVersion: v1
kind: ConfigMap
metadata:
  name: pod-names
  namespace: dispensary-scraper
data:
  names: |
    Aethelgard
    Xylos
    Kryll
    Coriolis
    Dimidium
    Veridia
    Zetani
    Talos IV
    Onyx
    Celestia
    Gormand
    Betha
    Ragnar
    Syphon
    Axiom
    Nadir
    Terra Nova
    Acheron
    Nexus
    Vespera
    Helios Prime
    Oasis
    Mordina
    Cygnus
    Umbra
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: worker-pod
  namespace: dispensary-scraper
spec:
  serviceName: worker-pods
  replicas: 5
  podManagementPolicy: Parallel
  selector:
    matchLabels:
      app: worker-pod
  template:
    metadata:
      labels:
        app: worker-pod
    spec:
      imagePullSecrets:
        - name: regcred
      containers:
        - name: workers
          image: code.cannabrands.app/creationshop/dispensary-scraper:latest
          # Run 5 workers per pod
          command: ["/bin/sh", "-c"]
          args:
            - |
              # Get pod ordinal (0, 1, 2, etc.)
              ORDINAL=$(echo $HOSTNAME | rev | cut -d'-' -f1 | rev)
              # Get pod name from configmap
              POD_NAME=$(sed -n "$((ORDINAL + 1))p" /etc/pod-names/names)
              echo "Starting pod: $POD_NAME (ordinal: $ORDINAL)"

              # Start 5 workers in this pod
              for i in 1 2 3 4 5; do
                WORKER_ID="${POD_NAME}-worker-${i}" \
                POD_NAME="$POD_NAME" \
                node dist/tasks/task-worker.js &
              done

              # Wait for all workers
              wait
          envFrom:
            - configMapRef:
                name: scraper-config
            - secretRef:
                name: scraper-secrets
          env:
            - name: API_BASE_URL
              value: "http://scraper:3010"
            - name: WORKERS_PER_POD
              value: "5"
          volumeMounts:
            - name: pod-names
              mountPath: /etc/pod-names
          resources:
            requests:
              memory: "512Mi"
              cpu: "200m"
            limits:
              memory: "1Gi"
              cpu: "1000m"
          livenessProbe:
            exec:
              command:
                - /bin/sh
                - -c
                - "pgrep -f 'task-worker' > /dev/null"
            initialDelaySeconds: 15
            periodSeconds: 30
            failureThreshold: 3
      volumes:
        - name: pod-names
          configMap:
            name: pod-names
      terminationGracePeriodSeconds: 60
---
# Headless service for StatefulSet
apiVersion: v1
kind: Service
metadata:
  name: worker-pods
  namespace: dispensary-scraper
spec:
  clusterIP: None
  selector:
    app: worker-pod
  ports:
    - port: 80
      name: placeholder
---
# HPA to scale pods based on pending tasks
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: worker-pod-hpa
  namespace: dispensary-scraper
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: StatefulSet
    name: worker-pod
  minReplicas: 5
  maxReplicas: 15
  metrics:
    - type: External
      external:
        metric:
          name: pending_tasks
          selector:
            matchLabels:
              queue: worker_tasks
        target:
          type: AverageValue
          averageValue: "10"