predictor/DEPLOYMENT.md

# Deployment Guide

This guide covers deploying the Predictor Service using Docker and Docker Compose.

## Prerequisites

- Docker Engine 20.10+
- Docker Compose 2.0+
- At least 2GB RAM available
- 10GB free disk space

## Quick Deployment

### 1. Clone and Setup

```bash
git clone <repository-url>
cd predictor
```

### 2. Validate Configuration

```bash
# Validate Docker configuration
./scripts/validate-docker.sh
```

### 3. Deploy

```bash
# Build and start services
make up-build

# Check status
make ps

# View logs
make logs
```

## Production Deployment

### Environment Configuration

1. **Copy environment template:**
   ```bash
   cp cmd/api/.env cmd/api/.env.production
   ```

2. **Edit production environment:**
   ```bash
   nano cmd/api/.env.production
   ```

3. **Key production settings:**
   ```bash
   # Security
   GSN_PREDICTOR_REDIS_PASSWORD=your_secure_password
   
   # Performance
   GSN_PREDICTOR_GRIB_PARALLEL=8
   GSN_PREDICTOR_GRIB_CACHE_TTL=2h
   
   # Monitoring
   GSN_PREDICTOR_GRIB_UPDATER_INTERVAL=3h
   ```

### Production Docker Compose

Create `docker-compose.prod.yml`:

```yaml
version: '3.8'

services:
  predictor:
    build:
      context: .
      dockerfile: Dockerfile
    container_name: predictor-prod
    ports:
      - "8080:8080"
    env_file:
      - cmd/api/.env.production
    volumes:
      - grib_data:/tmp/grib
    depends_on:
      redis:
        condition: service_healthy
    networks:
      - predictor-network
    restart: unless-stopped
    deploy:
      resources:
        limits:
          memory: 1G
          cpus: '0.5'
        reservations:
          memory: 512M
          cpus: '0.25'
    healthcheck:
      test: ["CMD", "wget", "--no-verbose", "--tries=1", "--spider", "http://localhost:8080/health"]
      interval: 30s
      timeout: 10s
      retries: 3
      start_period: 40s

  redis:
    image: redis:7.2-alpine
    container_name: predictor-redis-prod
    ports:
      - "6379:6379"
    volumes:
      - redis_data:/data
    networks:
      - predictor-network
    restart: unless-stopped
    command: redis-server --appendonly yes --maxmemory 512mb --maxmemory-policy allkeys-lru --requirepass ${GSN_PREDICTOR_REDIS_PASSWORD}
    healthcheck:
      test: ["CMD", "redis-cli", "-a", "${GSN_PREDICTOR_REDIS_PASSWORD}", "ping"]
      interval: 10s
      timeout: 3s
      retries: 5
      start_period: 10s

volumes:
  grib_data:
    driver: local
  redis_data:
    driver: local

networks:
  predictor-network:
    driver: bridge
```

### Deploy to Production

```bash
# Deploy with production config
docker-compose -f docker-compose.prod.yml up -d

# Monitor deployment
docker-compose -f docker-compose.prod.yml logs -f

# Check health
curl http://localhost:8080/health
```

## Kubernetes Deployment

### Create Namespace

```yaml
# k8s/namespace.yaml
apiVersion: v1
kind: Namespace
metadata:
  name: predictor
```

### Redis Deployment

```yaml
# k8s/redis.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: redis
  namespace: predictor
spec:
  replicas: 1
  selector:
    matchLabels:
      app: redis
  template:
    metadata:
      labels:
        app: redis
    spec:
      containers:
      - name: redis
        image: redis:7.2-alpine
        ports:
        - containerPort: 6379
        command: ["redis-server", "--appendonly", "yes", "--maxmemory", "512mb", "--maxmemory-policy", "allkeys-lru"]
        volumeMounts:
        - name: redis-data
          mountPath: /data
        resources:
          limits:
            memory: "512Mi"
            cpu: "250m"
          requests:
            memory: "256Mi"
            cpu: "100m"
        livenessProbe:
          exec:
            command: ["redis-cli", "ping"]
          initialDelaySeconds: 10
          periodSeconds: 10
        readinessProbe:
          exec:
            command: ["redis-cli", "ping"]
          initialDelaySeconds: 5
          periodSeconds: 5
      volumes:
      - name: redis-data
        persistentVolumeClaim:
          claimName: redis-pvc
---
apiVersion: v1
kind: Service
metadata:
  name: redis
  namespace: predictor
spec:
  selector:
    app: redis
  ports:
  - port: 6379
    targetPort: 6379
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: redis-pvc
  namespace: predictor
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 10Gi
```

### Predictor Deployment

```yaml
# k8s/predictor.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: predictor
  namespace: predictor
spec:
  replicas: 2
  selector:
    matchLabels:
      app: predictor
  template:
    metadata:
      labels:
        app: predictor
    spec:
      containers:
      - name: predictor
        image: predictor:latest
        ports:
        - containerPort: 8080
        env:
        - name: GSN_PREDICTOR_REDIS_HOST
          value: "redis"
        - name: GSN_PREDICTOR_REDIS_PORT
          value: "6379"
        - name: GSN_PREDICTOR_GRIB_DIR
          value: "/tmp/grib"
        - name: GSN_PREDICTOR_SCHEDULER_ENABLED
          value: "true"
        - name: GSN_PREDICTOR_GRIB_UPDATER_INTERVAL
          value: "6h"
        - name: GSN_PREDICTOR_GRIB_UPDATER_TIMEOUT
          value: "45m"
        volumeMounts:
        - name: grib-data
          mountPath: /tmp/grib
        resources:
          limits:
            memory: "1Gi"
            cpu: "500m"
          requests:
            memory: "512Mi"
            cpu: "250m"
        livenessProbe:
          httpGet:
            path: /health
            port: 8080
          initialDelaySeconds: 40
          periodSeconds: 30
        readinessProbe:
          httpGet:
            path: /health
            port: 8080
          initialDelaySeconds: 10
          periodSeconds: 10
      volumes:
      - name: grib-data
        emptyDir: {}
---
apiVersion: v1
kind: Service
metadata:
  name: predictor
  namespace: predictor
spec:
  selector:
    app: predictor
  ports:
  - port: 80
    targetPort: 8080
  type: LoadBalancer
```

### Deploy to Kubernetes

```bash
# Apply namespace
kubectl apply -f k8s/namespace.yaml

# Apply Redis
kubectl apply -f k8s/redis.yaml

# Wait for Redis to be ready
kubectl wait --for=condition=ready pod -l app=redis -n predictor

# Apply Predictor
kubectl apply -f k8s/predictor.yaml

# Check status
kubectl get pods -n predictor
kubectl get services -n predictor
```

## Monitoring and Logging

### Health Checks

The service includes built-in health checks:

```bash
# Application health
curl http://localhost:8080/health

# Docker health
docker inspect predictor | jq '.[0].State.Health'

# Kubernetes health
kubectl describe pod -l app=predictor -n predictor
```

### Logging

```bash
# Docker logs
docker-compose logs -f predictor

# Kubernetes logs
kubectl logs -f deployment/predictor -n predictor
```

### Metrics

Consider adding Prometheus metrics:

```yaml
# Add to docker-compose.yml
  prometheus:
    image: prom/prometheus:latest
    ports:
      - "9090:9090"
    volumes:
      - ./monitoring/prometheus.yml:/etc/prometheus/prometheus.yml
    networks:
      - predictor-network
```

## Backup and Recovery

### Redis Backup

```bash
# Create backup
docker exec predictor-redis redis-cli BGSAVE

# Copy backup file
docker cp predictor-redis:/data/dump.rdb ./backup/redis-$(date +%Y%m%d).rdb
```

### GRIB Data Backup

```bash
# Backup GRIB data
docker run --rm -v predictor_grib_data:/data -v $(pwd)/backup:/backup alpine tar czf /backup/grib-$(date +%Y%m%d).tar.gz -C /data .
```

### Automated Backup Script

```bash
#!/bin/bash
# scripts/backup.sh

BACKUP_DIR="./backup/$(date +%Y%m%d)"
mkdir -p $BACKUP_DIR

# Redis backup
docker exec predictor-redis redis-cli BGSAVE
sleep 5
docker cp predictor-redis:/data/dump.rdb $BACKUP_DIR/redis.rdb

# GRIB data backup
docker run --rm -v predictor_grib_data:/data -v $(pwd)/$BACKUP_DIR:/backup alpine tar czf /backup/grib.tar.gz -C /data .

echo "Backup completed: $BACKUP_DIR"
```

## Troubleshooting

### Common Issues

1. **Redis Connection Issues:**
   ```bash
   # Check Redis status
   docker-compose exec redis redis-cli ping
   
   # Check network connectivity
   docker-compose exec predictor wget -O- http://redis:6379
   ```

2. **GRIB Download Failures:**
   ```bash
   # Check disk space
   docker-compose exec predictor df -h /tmp/grib
   
   # Check internet connectivity
   docker-compose exec predictor wget -O- https://nomads.ncep.noaa.gov/
   ```

3. **Memory Issues:**
   ```bash
   # Check memory usage
   docker stats
   
   # Check container logs
   docker-compose logs predictor | grep -i memory
   ```

### Performance Tuning

1. **Redis Optimization:**
   ```bash
   # Increase Redis memory
   GSN_PREDICTOR_REDIS_MAXMEMORY=1gb
   
   # Optimize Redis settings
   redis-server --maxmemory 1gb --maxmemory-policy allkeys-lru
   ```

2. **GRIB Processing:**
   ```bash
   # Increase parallel workers
   GSN_PREDICTOR_GRIB_PARALLEL=8
   
   # Optimize cache TTL
   GSN_PREDICTOR_GRIB_CACHE_TTL=2h
   ```

3. **Container Resources:**
   ```yaml
   # In docker-compose.yml
   deploy:
     resources:
       limits:
         memory: 2G
         cpus: '1.0'
       reservations:
         memory: 1G
         cpus: '0.5'
   ```

## Security Considerations

1. **Network Security:**
   - Use internal networks for service communication
   - Expose only necessary ports
   - Use reverse proxy for external access

2. **Container Security:**
   - Run as non-root user
   - Use minimal base images
   - Regular security updates

3. **Data Security:**
   - Encrypt sensitive environment variables
   - Use secrets management for passwords
   - Regular backups

4. **Access Control:**
   - Implement API authentication
   - Use HTTPS in production
   - Monitor access logs
```