How do you implement Docker container resource quotas and limits in production?

Docker container resource quotas and limits for production: 1) Memory limits: Hard limits: docker run -m 1g myapp (container killed if exceeded), Soft limits: --memory-reservation 512m (reclaimed under pressure), Swap control: --memory-swap 2g (total memory + swap), OOM behavior: --oom-kill-disable (prevent OOM killer), 2) CPU limits: CPU shares: --cpu-shares 512 (relative weight), CPU quota: --cpus 1.5 (absolute limit), CPU affinity: --cpuset-cpus 0,1 (specific cores), 3) I/O limits: Block I/O: --device-read-bps /dev/sda:1mb, --device-write-bps /dev/sda:1mb, --blkio-weight 500 (relative I/O priority), 4) Network limits: Bandwidth limiting with tc (traffic control), Container network namespaces, 5) Production implementation: Docker Compose deploy resources, Kubernetes resource requests/limits, Monitoring with cAdvisor, Prometheus metrics, 6) Best practices: Set both requests and limits, Monitor resource usage, Implement alerting, Test under load, Plan for peak usage, Use horizontal scaling. Example: docker run --memory 1g --cpus 1.0 --device-read-bps /dev/sda:10mb myapp.