Question
📝 **Question:** Why use a multi-stage Dockerfile for Python services?
📋 Pick the right answer.
💡 **Hint:** Re-read the theory above if unsure.
A It's the only Dockerfile shape modern BuildKit accepts — single-stage Dockerfiles silently fail on `docker build` since the 24.0 release that landed across the major Linux distributions late last year B Build tools (gcc, dev headers, uv) stay in the builder stage; only compiled wheels and source make it to the runtime image. Result: 5-10x smaller production images, faster pulls, faster cold starts C Kubernetes requires it as of v1.30 — the kubelet refuses to schedule pods built from a single-stage image because the layer-cache hashing changed for OCI v1.1 spec compliance reasons across the cluster D Multi-stage images run noticeably faster at request-handling time because the Python interpreter walks fewer layer-overlay paths during import resolution, especially under high concurrency and cold-cache loads