Production Deployment Guide

Pre-deployment checklist

• Configure an authentication password and, if needed, TLS.
• Set memory limits and eviction policy.
• Define connection limits and timeouts.
• Enable HTTP health checks for load balancers.
• Run load tests with production-like traffic.
• Set up monitoring and alerts.
• Define a rollback and cache-warming strategy.

Essential configuration

Authentication

[security]
password = "your-secure-password"

Or via environment variable:

export REDIS_PASSWORD="your-secure-password"
./redistill

Memory limits

[memory]
max_memory = 2147483648  # 2GB
eviction_policy = "allkeys-lru"

Connection management

[server]
max_connections = 10000
connection_rate_limit = 1000
connection_timeout = 300

Health checks

[server]
health_check_port = 8080

Health endpoint: http://<host>:8080/health

TLS/SSL

[security]
tls_enabled = true
tls_cert_path = "/etc/letsencrypt/live/your-domain.com/fullchain.pem"
tls_key_path = "/etc/letsencrypt/live/your-domain.com/privkey.pem"

Deployment options

Docker

docker run -d --name redistill \
  -p 6379:6379 -p 8080:8080 \
  -e REDIS_PASSWORD=$REDIS_PASSWORD \
  -v ./redistill.toml:/etc/redistill/redistill.toml:ro \
  shahidontech/redistill:latest

systemd

Example service file:

[Unit]
Description=Redistill Cache Server
After=network-online.target

[Service]
Type=simple
User=redistill
WorkingDirectory=/opt/redistill
EnvironmentFile=/etc/redistill/environment
ExecStart=/opt/redistill/redistill
Restart=on-failure
LimitNOFILE=65536

[Install]
WantedBy=multi-user.target

Kubernetes

Run Redistill as a Deployment with a ClusterIP Service and HTTP health probes:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: redistill
spec:
  replicas: 3
  selector:
    matchLabels:
      app: redistill
  template:
    metadata:
      labels:
        app: redistill
    spec:
      containers:
      - name: redistill
        image: your-registry/redistill:latest
        ports:
        - containerPort: 6379
        - containerPort: 8080
        env:
        - name: REDIS_PASSWORD
          valueFrom:
            secretKeyRef:
              name: redistill-secret
              key: password
        livenessProbe:
          httpGet:
            path: /health
            port: 8080
        readinessProbe:
          httpGet:
            path: /health
            port: 8080

Monitoring & alerting

Key metrics to track:

• Memory: used_memory vs max_memory
• Evictions: evicted_keys rate
• Connections: active_connections, rejected_connections
• Throughput: total_commands over time

All of these are exposed via INFO and the health endpoint.

Scaling strategies

Vertical scaling

• Add more CPU cores and RAM.
• Increase num_shards to match CPU count.

Horizontal scaling

• Client-side sharding with consistent hashing.
• Proxy-based sharding using Twemproxy, Envoy or HAProxy.
• DNS round-robin for simple read-heavy workloads.

Persistence & cache warming

By default, persistence is disabled for maximum performance and Redistill behaves like a pure cache. When enabling snapshots, be aware that data between snapshots can be lost on crash.

• Use cache warming to pre-load the most important keys on startup.
• Ensure your application has a well-defined fallback to the primary data store on cache miss.

Security hardening

• Run Redistill under a dedicated system user with restricted filesystem permissions.
• Bind to private networks and use firewalls/security groups to limit access.
• Enable TLS for external traffic and rotate credentials regularly.
• Monitor authentication failures and connection rejections as part of your security posture.