14.1 Production Readiness

Production readiness is the overall assessment of whether a containerized application has actually addressed the full range of concerns — resource limits, health checks, logging, security, graceful shutdown, monitoring — necessary to run reliably and safely under real production conditions, rather than simply having been confirmed to work in development.

A Representative Production Readiness Checklist

Walking through a specific, concrete set of considerations helps surface gaps before they become real incidents.

services:
  app:
    image: registry.example.com/myapp:1.0
    deploy:
      resources:
        limits:
          memory: 512M
      restart_policy:
        condition: on-failure
        max_attempts: 3
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8080/health"]
    logging:
      driver: json-file
      options:
        max-size: "10m"

This single configuration addresses several distinct readiness concerns at once: resource constraints, restart behavior, health monitoring, and bounded log retention.

Why Each Missing Piece Represents a Specific, Real Risk

Skipping resource limits risks one runaway container affecting others; skipping health checks risks an unhealthy container continuing to receive traffic; skipping bounded logging risks disk exhaustion over time.

docker stats
docker inspect myapp --format '{{.State.Health.Status}}'

Verifying Graceful Shutdown Behavior

Confirming the application correctly handles a termination signal, finishing in-flight work before exiting, avoids dropped requests during a routine restart or scaling event.

docker stop --time=30 myapp

Confirming Observability Is Actually in Place

Verifying that logs, metrics, and any relevant alerting are genuinely configured and functioning — not just assumed to be — is itself part of a thorough readiness assessment.

docker logs myapp --tail 50

Why Production Readiness Matters

Working through a concrete, comprehensive checklist before considering a deployment genuinely production-ready surfaces specific, addressable gaps in advance, considerably preferable to discovering them only after a real production incident has already occurred.