本文へ移動
cccskills
無料GitHub で公開

optimize-docker-build-cache

Optimize Docker build times using layer caching, multi-stage builds, BuildKit features, and dependency-first copy patterns. Applicable to R, Node.js, and Python projects. Use when Docker builds are slow due to repeated package installations, when rebuilds reinstall all dependencies on every code change, when image sizes are unnecessarily large, or when CI/CD pipeline builds are a bottleneck.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md8.2 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Optimize Docker Build Cache

Reduce Docker build times through effective layer caching and build optimization.

When to Use

  • Docker builds are slow due to repeated package installations
  • Rebuilds reinstall all dependencies on every code change
  • Image sizes are unnecessarily large
  • CI/CD pipeline builds are a bottleneck

Inputs

  • Required: Existing Dockerfile to optimize
  • Optional: Target build time improvement
  • Optional: Target image size reduction

Procedure

Step 1: Order Layers by Change Frequency

Place least-changing layers first:

# 1. Base image (rarely changes)
FROM rocker/r-ver:4.5.0

# 2. System dependencies (change occasionally)
RUN apt-get update && apt-get install -y \
    libcurl4-openssl-dev \
    libssl-dev \
    && rm -rf /var/lib/apt/lists/*

# 3. Dependency files only (change when deps change)
COPY renv.lock renv.lock
COPY renv/activate.R renv/activate.R
RUN R -e "renv::restore()"

# 4. Source code (changes frequently)
COPY . .

Key principle: Docker caches each layer. When a layer changes, all subsequent layers are rebuilt. Dependency installation should come before source code copy.

Expected: The Dockerfile layers are ordered from least-changing (base image, system deps) to most-changing (source code), with dependency lockfiles copied before the full source.

On failure: If builds still reinstall dependencies on every code change, verify that COPY . . comes after the dependency installation RUN command, not before.

Step 2: Separate Dependency Installation from Code

Bad (rebuilds packages on every code change):

COPY . .
RUN R -e "renv::restore()"

Good (only rebuilds packages when lockfile changes):

COPY renv.lock renv.lock
RUN R -e "renv::restore()"
COPY . .

Same pattern for Node.js:

COPY package.json package-lock.json ./
RUN npm ci
COPY . .

Expected: Dependency lockfile (renv.lock, package-lock.json, requirements.txt) is copied and installed in a separate layer before the full source code COPY . ..

On failure: If the lockfile copy fails, ensure the file exists in the build context and is not excluded by .dockerignore.

Step 3: Use Multi-Stage Builds

Separate build dependencies from runtime:

# Build stage - includes dev tools
FROM rocker/r-ver:4.5.0 AS builder
RUN apt-get update && apt-get install -y \
    libcurl4-openssl-dev libssl-dev build-essential
COPY renv.lock .
RUN R -e "install.packages('renv'); renv::restore()"

# Runtime stage - minimal image
FROM rocker/r-ver:4.5.0
RUN apt-get update && apt-get install -y \
    libcurl4 libssl3 \
    && rm -rf /var/lib/apt/lists/*
COPY --from=builder /usr/local/lib/R/site-library /usr/local/lib/R/site-library
COPY . /app
WORKDIR /app
CMD ["Rscript", "main.R"]

Expected: The Dockerfile has a builder stage with dev tools and a runtime stage with only production dependencies. The final image is significantly smaller than a single-stage build.

On failure: If COPY --from=builder fails to find libraries, verify the install path matches between stages. Use docker build --target builder . to debug the build stage independently.

Step 4: Combine RUN Commands

Each RUN creates a layer. Combine related commands:

Bad (3 layers, apt cache persists):

RUN apt-get update
RUN apt-get install -y curl git
RUN rm -rf /var/lib/apt/lists/*

Good (1 layer, clean cache):

RUN apt-get update && apt-get install -y \
    curl \
    git \
    && rm -rf /var/lib/apt/lists/*

Expected: Related apt-get or package install commands are combined into single RUN instructions, each ending with cache cleanup (rm -rf /var/lib/apt/lists/*).

On failure: If a combined RUN command fails midway, temporarily split it to identify the failing command, then recombine after fixing.

Step 5: Use .dockerignore

Prevent unnecessary files from entering the build context:

.git
.Rproj.user
.Rhistory
.RData
renv/library
renv/cache
node_modules
docs/
*.tar.gz
.env

Expected: A .dockerignore file exists in the project root excluding .git, node_modules, renv/library, build artifacts, and environment files. Build context size is noticeably smaller.

On failure: If needed files are missing in the container, check .dockerignore for overly broad patterns. Use docker build verbose output to verify which files are sent to the daemon.

Step 6: Enable BuildKit

DOCKER_BUILDKIT=1 docker build -t myimage .

Or in docker-compose.yml:

services:
  app:
    build:
      context: .
      dockerfile: Dockerfile

With COMPOSE_DOCKER_CLI_BUILD=1 and DOCKER_BUILDKIT=1 environment variables.

BuildKit enables:

  • Parallel stage builds
  • Better cache management
  • --mount=type=cache for persistent package caches

Expected: Builds run with BuildKit enabled (indicated by #1 [internal] load build definition style output). Multi-stage builds execute stages in parallel where possible.

On failure: If BuildKit is not active, verify the environment variables are exported before the build command. On older Docker versions, upgrade Docker Engine to 18.09+ for BuildKit support.

Step 7: Use Cache Mounts for Package Managers

# R packages with persistent cache
RUN --mount=type=cache,target=/usr/local/lib/R/site-library \
    R -e "install.packages('dplyr')"

# npm with persistent cache
RUN --mount=type=cache,target=/root/.npm \
    npm ci

Expected: Subsequent builds reuse cached packages from the mount, dramatically reducing install times even when the layer is invalidated. Cache persists across builds.

On failure: If --mount=type=cache is not recognized, ensure BuildKit is enabled (DOCKER_BUILDKIT=1). The syntax requires BuildKit and is not supported by the legacy builder.

Validation

  • Rebuilds after code-only changes are significantly faster
  • Dependency installation layer is cached when lockfile hasn't changed
  • .dockerignore excludes unnecessary files
  • Image size is reduced compared to unoptimized build
  • Multi-stage build (if used) separates build and runtime dependencies

Common Pitfalls

  • Copying all files before installing deps: Invalidates the dependency cache on every code change
  • Forgetting .dockerignore: Large build contexts slow down every build
  • Too many layers: Each RUN, COPY, ADD creates a layer. Combine where logical.
  • Not cleaning apt cache: Always end apt-get installs with && rm -rf /var/lib/apt/lists/*
  • Platform-specific caches: Cache layers are platform-specific. CI runners may not benefit from local caches.
  • No cache between CI runs: Ephemeral runners start with an empty local cache, so every run rebuilds from scratch. Export the cache to a registry with --cache-to type=registry,ref=<registry>/<cache-image>,mode=max and import it on the next run with --cache-from type=registry,ref=<registry>/<cache-image>; mode=max also caches layers from intermediate stages, where the default mode=min caches only layers present in the final image. Neither this backend nor the GitHub Actions one (type=gha) is available on the default docker driver without extra setup (type=registry requires the containerd image store enabled), so create a dedicated builder first with docker buildx create --driver docker-container --use.

Related Skills

  • create-r-dockerfile - initial Dockerfile creation
  • setup-docker-compose - compose build configuration
  • containerize-mcp-server - apply optimizations to MCP server builds

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Execute structural metamorphosis using strangler fig migration, chrysalis phases, and interface preservation. Covers transformation planning, parallel running, progressive cutover, rollback design, and post-metamorphosis stabilization for system architecture evolution. Use when assess-form has classified the system as READY for transformation, when migrating from monolith to microservices, when replacing a core subsystem while dependents continue operating, or when any architectural change must be gradual rather than big-bang.

日本語の概要は準備中です。原文の説明を表示しています。

pjt222/agent-almanac372026年10月10日 更新

adaptic

無料

Master skill composing the 5-step synoptic cycle for panoramic synthesis across multiple domains. Orchestrates meditate, expand-awareness, observe, awareness, integrate-gestalt, and express-insight into a coherent process that produces unified understanding rather than sequential compromise. Use when a problem genuinely spans 3+ domains and the interactions between domains matter more than depth in any one, when sequential analysis feels like compromise rather than integration, or before major architectural decisions affecting multiple stakeholders.

日本語の概要は準備中です。原文の説明を表示しています。

pjt222/agent-almanac372026年10月10日 更新

Scaffold a new puzzle type across all 10+ pipeline integration points in jigsawR. Creates the core puzzle module, wires it into the unified pipeline (generation, positioning, rendering, adjacency), adds ggpuzzle geom/stat layers, updates DESCRIPTION and config.yml, extends the Shiny app, and creates a comprehensive test suite. Use when adding a completely new puzzle type to the package or following the 10-point integration checklist to ensure nothing is missed end-to-end.

日本語の概要は準備中です。原文の説明を表示しています。

pjt222/agent-almanac372026年10月10日 更新

Add Rcpp or RcppArmadillo integration to an R package for high-performance C++ code. Covers setup, writing C++ functions, RcppExports generation, testing compiled code, and debugging. Use when an R function is too slow and profiling confirms a bottleneck, when you need to interface with existing C/C++ libraries, or when implementing algorithms (loops, recursion, linear algebra) that benefit from compiled code.

日本語の概要は準備中です。原文の説明を表示しています。

pjt222/agent-almanac372026年10月10日 更新

aikido

無料

Practice aikido as a defensive martial art emphasizing harmony, redirection, and controlled resolution. Covers centering and grounding, ukemi (safe falling and rolling), core techniques (irimi-nage, shiho-nage, kote-gaeshi, ikkyo), blending and tai sabaki (body movement), randori (multiple attacker practice), weapons awareness, and applying principles off the mat. Use when learning a defensive martial art that prioritizes de-escalation, developing the ability to redirect aggression without causing unnecessary harm, building safe falling skills, or cultivating calm centeredness under physical pressure.

日本語の概要は準備中です。原文の説明を表示しています。

pjt222/agent-almanac372026年10月10日 更新

Analyze an arbitrary codebase to identify functions, APIs, and data sources suitable for exposure as MCP tools, producing a tool specification document. Use when planning an MCP server for an existing project, auditing a codebase before wrapping it as an AI-accessible tool surface, comparing what a codebase can do versus what is already exposed via MCP, or generating a tool spec to hand off to scaffold-mcp-server.

日本語の概要は準備中です。原文の説明を表示しています。

pjt222/agent-almanac372026年10月10日 更新

pjt222 のスキルをすべて見る

このスキルの問題を報告する