Where Docker Ends

Single Host → Fleet

Everything in this track manages containers on one machine. Production at scale adds requirements a single daemon can't meet:

Need	Single-host answer	Fleet answer
App composed of services	docker compose	Kubernetes Deployments/Services
Restart on failure	`--restart`	self-healing ReplicaSets
More capacity	bigger VM	scheduling across nodes
Zero-downtime deploys	manual blue/green	rolling updates built-in
Service discovery	user-defined network DNS	cluster DNS + Services
Secrets	file mounts	Secrets objects + RBAC

The crucial insight: Kubernetes doesn't replace what you learned - it consumes it. Your images, Dockerfiles, health endpoints, resource limits, non-root users and secret files are exactly what k8s manifests reference. A pod spec is your docker run flags, written as YAML, scheduled across a fleet.

# The docker run you know…        # …is the pod spec you'll write
# --memory 256m                   resources: { limits: { memory: 256Mi } }
# --health-cmd …                  livenessProbe: { httpGet: … }
# --user 1000                     securityContext: { runAsUser: 1000 }
# -v secret:/run/secrets/x:ro     volumes: secret: …

Note: Swarm (docker swarm init) still exists and is genuinely simple for small fleets, but the industry consolidated on Kubernetes - which is the next track on this platform.

Goal: You've covered the entire Docker surface: fundamentals, images, data, networking, compose, debugging, observability, CI/CD, security and production practice - with 30 hands-on incident labs. Finish any labs you skipped, then you're ready for the Kubernetes track when it lands.