LOCAL PREVIEW View on GitHub

Docker Interview Pack

This folder contains Docker-focused interview material grounded in the MangaAssist project.

What This Covers

  • ECS Fargate as the steady-state container platform
  • ECR, multi-stage Docker builds, and image minimization
  • SageMaker inference containers for custom model serving
  • vLLM container optimizations for throughput and cost
  • Container startup, warmup, and cold-start mitigation
  • Dockerized integration testing with LocalStack and TestContainers
  • Container supply-chain security with scans, SBOMs, signatures, and rollback policy

Primary Study File

Companion LLD

Individual Scenario Deep Dives

Each scenario broken out separately with: full story grounded to MangaAssist, deep-dive Q&A, decision tables, tradeoffs discussed, scale planned, and intuition gained.

# File Topic
1 scenario-01-multistage-fargate.md Multi-stage builds + ECS Fargate hybrid compute
2 scenario-02-coldstart-elimination.md SageMaker inference container cold-start elimination
3 scenario-03-vllm-serving.md vLLM serving container — throughput, cost, operability
4 scenario-04-gpu-oom-containment.md GPU OOM → container restart containment
5 scenario-05-dockerized-ci-testing.md Dockerized CI integration testing (LocalStack + TestContainers + WireMock)
6 scenario-06-supply-chain-security.md Container supply-chain security and release gating

Grounding Documents