Performance Optimization¶

Guidelines for optimizing Omakase homelab performance.

Resource Allocation¶

CPU Limits¶

Set appropriate CPU limits for each service:

deploy:
  resources:
    limits:
      cpus: '2.0'      # Maximum 2 CPU cores
    reservations:
      cpus: '0.5'      # Minimum 0.5 CPU cores

Recommended allocations: - Lightweight services (Traefik, Authelia): 0.5-1 CPU - Medium services (Nextcloud, Jellyfin): 1-2 CPUs - Heavy services (Local AI, Immich): 2-4 CPUs

Memory Limits¶

Prevent memory exhaustion:

deploy:
  resources:
    limits:
      memory: 2G       # Maximum 2GB RAM
    reservations:
      memory: 512M     # Minimum 512MB RAM

Recommended allocations: - Lightweight services: 256M-512M - Medium services: 512M-2G - Heavy services: 2G-8G - Databases: 1G-4G

Monitor Resource Usage¶

# Real-time monitoring
docker stats

# Identify resource hogs
docker stats --no-stream | sort -k 3 -h

Network Performance¶

Network Isolation¶

Use dedicated networks per service to reduce broadcast traffic:

networks:
  vnet-service:
    driver: bridge
    ipam:
      config:
        - subnet: 192.168.X.0/24

Traefik Optimization¶

Enable HTTP/2:

entryPoints:
  websecure:
    http2:
      maxConcurrentStreams: 250

Enable compression:

http:
  middlewares:
    compression:
      compress: {}

Connection limits:

entryPoints:
  websecure:
    transport:
      respondingTimeouts:
        readTimeout: 60s
        writeTimeout: 60s

Storage Performance¶

Volume Configuration¶

Use named volumes for databases:

volumes:
  postgres_data:
    driver: local
    driver_opts:
      type: none
      o: bind
      device: ${DATA_DIR}/postgres/data

Database Optimization¶

PostgreSQL:

environment:
  POSTGRES_INITDB_ARGS: "-E UTF8 --locale=C"
  # Tune for your RAM
  POSTGRES_SHARED_BUFFERS: "256MB"
  POSTGRES_EFFECTIVE_CACHE_SIZE: "1GB"
  POSTGRES_MAX_CONNECTIONS: "100"

Redict (Redis alternative):

command: >
  --save 60 1000
  --maxmemory 256mb
  --maxmemory-policy allkeys-lru

Disk I/O¶

Monitor disk performance:

iostat -x 1

Use SSD for databases - Store database volumes on SSD for best performance.

Mount options - Use noatime to reduce write operations:

mount -o remount,noatime ${DATA_DIR}

Application Optimization¶

Caching¶

Enable application-level caching where available:

Nextcloud:

environment:
  REDIS_HOST: redict
  REDIS_HOST_PORT: 6379

Immich:

environment:
  IMMICH_MACHINE_LEARNING_ENABLED: "true"
  IMMICH_MACHINE_LEARNING_URL: http://immich-ml:3003

Concurrent Connections¶

Tune application workers:

Gunicorn-based apps:

environment:
  GUNICORN_WORKERS: 4
  GUNICORN_THREADS: 2

Node.js apps:

environment:
  NODE_OPTIONS: "--max-old-space-size=2048"
  UV_THREADPOOL_SIZE: 4

Container Optimization¶

Multi-stage Builds¶

For custom images, use multi-stage builds to reduce image size.

Image Selection¶

Prefer Alpine-based images when possible: - Smaller size - Faster startup - Lower memory footprint

Restart Policies¶

Use appropriate restart policies:

restart: unless-stopped  # Recommended for most services
restart: always         # Critical services only
restart: on-failure     # Services that should stay down when intentionally stopped

Monitoring Performance¶

Identify Bottlenecks¶

CPU bottlenecks:

docker stats --format "table {{.Name}}\t{{.CPUPerc}}" | sort -k 2 -h

Memory bottlenecks:

docker stats --format "table {{.Name}}\t{{.MemPerc}}" | sort -k 2 -h

I/O bottlenecks:

docker stats --format "table {{.Name}}\t{{.BlockIO}}"

Network bottlenecks:

docker stats --format "table {{.Name}}\t{{.NetIO}}"

Performance Metrics¶

Track over time: - Response times (Traefik dashboard) - Database query times - Container startup times - Backup durations

Scaling Strategies¶

Horizontal Scaling¶

For services that support it, deploy multiple instances:

deploy:
  replicas: 3

Vertical Scaling¶

Increase resources for resource-constrained services:

deploy:
  resources:
    limits:
      cpus: '4.0'
      memory: 8G

Load Balancing¶

Use Traefik's load balancing for scaled services:

labels:
  - traefik.http.services.myservice.loadbalancer.sticky.cookie=true

System-Level Optimization¶

Kernel Parameters¶

Optimize for containerized workloads:

# Increase file descriptors
ulimit -n 65536

# Optimize network
sysctl -w net.core.somaxconn=1024
sysctl -w net.ipv4.tcp_max_syn_backlog=2048

Swap Configuration¶

# Reduce swappiness for better performance
sysctl vm.swappiness=10

Docker Daemon¶

Optimize Docker daemon in /etc/docker/daemon.json:

{
  "log-driver": "json-file",
  "log-opts": {
    "max-size": "10m",
    "max-file": "3"
  },
  "storage-driver": "overlay2",
  "default-ulimits": {
    "nofile": {
      "Name": "nofile",
      "Hard": 64000,
      "Soft": 64000
    }
  }
}

Benchmarking¶

Web Service Response Time¶

# Using curl
curl -w "@curl-format.txt" -o /dev/null -s https://service.yourdomain.com

# curl-format.txt content:
#     time_namelookup:  %{time_namelookup}\n
#        time_connect:  %{time_connect}\n
#     time_appconnect:  %{time_appconnect}\n
#    time_pretransfer:  %{time_pretransfer}\n
#       time_redirect:  %{time_redirect}\n
#  time_starttransfer:  %{time_starttransfer}\n
#                     ----------\n
#          time_total:  %{time_total}\n

Database Performance¶

# PostgreSQL
docker exec postgres pgbench -U user -d dbname -c 10 -t 100

# Redict
docker exec redict redis-benchmark -h localhost -p 6379 -n 100000

Performance Checklist¶

Initial Setup: - [ ] Set resource limits for all services - [ ] Use dedicated networks for isolation - [ ] Enable caching where available - [ ] Use SSD for databases

Regular Optimization: - [ ] Monitor resource usage weekly - [ ] Identify and address bottlenecks - [ ] Review and adjust resource limits - [ ] Update to optimized image versions

Scaling: - [ ] Identify services that need more resources - [ ] Consider horizontal scaling for stateless services - [ ] Implement load balancing where needed - [ ] Monitor performance after scaling changes