Deployment Guide

This guide covers deploying Conduit in various production environments.

Deployment Options

| Option | Best For | | -------------- | ---------------------------------- | | Docker Compose | Small deployments, evaluation | | Kubernetes | Cloud-native, scalable deployments | | Air-Gapped | Secure, isolated environments |

Docker Compose Deployment

Basic Setup

# docker-compose.yaml
version: "3.8"

services:
  control-plane:
    image: conduit/control-plane:latest
    ports:
      - "3001:3001"
    environment:
      - DATABASE_URL=postgresql://conduit:password@postgres:5432/conduit
      - REDIS_URL=redis://redis:6379
      - NATS_URL=nats://nats:4222
      - LLM_PROVIDER=claude
      - ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY}
      - JWT_SECRET=${JWT_SECRET}
    depends_on:
      - postgres
      - redis
      - nats
    volumes:
      - ./config:/etc/conduit

  portal:
    image: conduit/portal:latest
    ports:
      - "5173:5173"
    environment:
      - VITE_API_URL=http://control-plane:3001

  postgres:
    image: pgvector/pgvector:pg15
    environment:
      - POSTGRES_USER=conduit
      - POSTGRES_PASSWORD=password
      - POSTGRES_DB=conduit
    volumes:
      - postgres-data:/var/lib/postgresql/data

  redis:
    image: redis:7-alpine
    ports:
      - "6379:6379"
    volumes:
      - redis-data:/data

  nats:
    image: nats:2.10-alpine
    command: >
      --name conduit-nats
      --js
      --sd /data
    ports:
      - "4222:4222"
      - "8222:8222"
    volumes:
      - nats-data:/data
    restart: unless-stopped

volumes:
  postgres-data:
  redis-data:
  nats-data:

Starting the Stack

# Start all services
docker-compose up -d

# View logs
docker-compose logs -f control-plane

# Check status
docker-compose ps

Kubernetes Deployment

Helm Chart

# Add Conduit Helm repository
helm repo add conduit https://charts.ulexioty.com
helm repo update

# Install Conduit
helm install conduit conduit/conduit \
  --namespace conduit \
  --create-namespace \
  --set controlPlane.replicas=3 \
  --set database.host=postgres.database.svc \
  --set nats.url=nats://nats.messaging.svc:4222

Custom Values

# values.yaml
controlPlane:
  replicas: 3
  image:
    repository: conduit/control-plane
    tag: latest

  resources:
    requests:
      cpu: 500m
      memory: 512Mi
    limits:
      cpu: 2000m
      memory: 2Gi

  ingress:
    enabled: true
    className: nginx
    hosts:
      - host: conduit.company.com
        paths:
          - path: /
            pathType: Prefix
    tls:
      - secretName: conduit-tls
        hosts:
          - conduit.company.com

database:
  host: postgres.database.svc
  port: 5432
  name: conduit
  username: conduit
  existingSecret: conduit-db-credentials

nats:
  url: nats://nats.messaging.svc:4222

monitoring:
  serviceMonitor:
    enabled: true

Manual Kubernetes Manifests

# deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: conduit-control-plane
  namespace: conduit
spec:
  replicas: 3
  selector:
    matchLabels:
      app: conduit-control-plane
  template:
    metadata:
      labels:
        app: conduit-control-plane
    spec:
      containers:
        - name: control-plane
          image: conduit/control-plane:latest
          ports:
            - containerPort: 8080
          env:
            - name: DATABASE_URL
              valueFrom:
                secretKeyRef:
                  name: conduit-secrets
                  key: database-url
          resources:
            requests:
              cpu: 500m
              memory: 512Mi
          livenessProbe:
            httpGet:
              path: /health
              port: 8080
            initialDelaySeconds: 30
          readinessProbe:
            httpGet:
              path: /ready
              port: 8080

Translator Deployment

Docker

docker run -d \
  --name conduit-translator-splunk \
  -e CONDUIT_NATS_URL=nats://nats:4222 \
  -e CONDUIT_API_KEY=${API_KEY} \
  -v /etc/conduit/adapters:/etc/conduit \
  conduit/translator-splunk:latest

Kubernetes DaemonSet

For deploying translators across nodes:

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: conduit-translator
  namespace: conduit-edge
spec:
  selector:
    matchLabels:
      app: conduit-translator
  template:
    spec:
      containers:
        - name: translator
          image: conduit/translator-splunk:latest
          env:
            - name: CONDUIT_NATS_URL
              value: nats://control-plane.conduit:4222
          volumeMounts:
            - name: config
              mountPath: /etc/conduit
      volumes:
        - name: config
          configMap:
            name: conduit-edge-config

Native Binary

For resource-constrained edge hardware:

# Download binary
curl -LO https://releases.ulexioty.com/translator/latest/conduit-translator-linux-arm64

# Install
chmod +x conduit-translator-linux-arm64
sudo mv conduit-translator-linux-arm64 /usr/local/bin/conduit-translator

# Create systemd service
sudo cat > /etc/systemd/system/conduit-translator.service << EOF
[Unit]
Description=Conduit Translator
After=network.target

[Service]
Type=simple
ExecStart=/usr/local/bin/conduit-translator --config /etc/conduit/translator.yaml
Restart=always
User=conduit

[Install]
WantedBy=multi-user.target
EOF

# Start service
sudo systemctl enable conduit-translator
sudo systemctl start conduit-translator

Air-Gapped Deployment

For environments without internet access:

1. Export Images

# On internet-connected machine
docker pull conduit/control-plane:latest
docker pull conduit/translator-splunk:latest
docker save conduit/control-plane:latest | gzip > conduit-control-plane.tar.gz
docker save conduit/translator-splunk:latest | gzip > conduit-translator-splunk.tar.gz

2. Transfer to Air-Gapped Environment

# Transfer files via approved media

3. Import Images

# On air-gapped machine
docker load < conduit-control-plane.tar.gz
docker load < conduit-translator-splunk.tar.gz

4. Configure for Offline

# config/conduit.yaml
updates:
  autoCheck: false

telemetry:
  enabled: false

# All dependencies must be internal
database:
  host: internal-postgres.local

nats:
  url: nats://internal-nats.local:4222

Production Checklist

Security

• [ ] TLS enabled for all connections
• [ ] Secrets managed via secret manager (Vault, AWS Secrets, etc.)
• [ ] Network policies restrict traffic
• [ ] RBAC configured with least privilege
• [ ] Audit logging enabled

High Availability

• [ ] Control Plane: 3+ replicas
• [ ] Database: Replicated (primary + standby)
• [ ] NATS: 3-node hub cluster
• [ ] Load balancer configured
• [ ] Health checks enabled

Monitoring

• [ ] Metrics exported (Prometheus)
• [ ] Dashboards created (Grafana)
• [ ] Alerts configured
• [ ] Log aggregation (ELK, Loki)
• [ ] Tracing enabled (Jaeger, Zipkin)

Backup & Recovery

• [ ] Database backups scheduled
• [ ] Configuration versioned in Git
• [ ] Disaster recovery plan documented
• [ ] Recovery procedures tested

Performance

• [ ] Resource limits set appropriately
• [ ] Connection pools sized correctly
• [ ] Query timeouts configured
• [ ] Caching enabled

Scaling Considerations

Horizontal Scaling

| Component | Scale Strategy | | ------------- | --------------------------------- | | Control Plane | Add replicas behind load balancer | | Translators | One per data source location | | Database | Read replicas for query scaling | | NATS | Hub cluster + leaf nodes |

Vertical Scaling

| Symptom | Solution | | ------------------- | --------------------------------- | | Slow queries | Increase Control Plane memory | | Connection timeouts | Increase pool sizes | | High latency | Add translators closer to sources |

Troubleshooting

Health Checks

# Control Plane health
curl https://conduit.company.com/health

# Translator status
curl http://localhost:9090/metrics | grep conduit_translator

Common Issues

Control Plane won't start

• Check database connectivity
• Verify NATS server is reachable
• Review logs: docker logs conduit-control-plane

Translator can't connect

• Verify API key is valid
• Check network path to Control Plane
• Confirm NATS ports are open (4222, 8222)

Slow query performance

• Check database indexes
• Review query patterns
• Consider adding caching

Next Steps

• Configuration Guide - Detailed configuration
• Architecture - System architecture
• API Reference - REST API docs