菜单
References cold start optimization history and techniques. Covers Docker image optimization, Fargate CPU tuning, SOCI lazy loading, and Lambda Bedrock discovery fix. Use when optimizing startup performance.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/serithemage/serverless-openclaw --skill cold-start命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
References Serverless OpenClaw architecture. Covers network design, CDK stack structure, DynamoDB data model, API protocols, and container design.
Loads Serverless OpenClaw project context. Provides background knowledge needed for development including project overview, tech stack, architecture decisions, and data models.
References Serverless OpenClaw cost optimization guidelines. Validates that no resources exceeding the cost target ($1/month) are created during implementation. Use when writing CDK stacks or making infrastructure changes.
基于 SOC 职业分类
正在显示 SKILL.md
| name | cold-start |
| description | References cold start optimization history and techniques. Covers Docker image optimization, Fargate CPU tuning, SOCI lazy loading, and Lambda Bedrock discovery fix. Use when optimizing startup performance. |
| allowed-tools | Read, Glob, Grep |
Full analysis: docs/cold-start-optimization.md
| Target | Before | After | Savings |
|---|---|---|---|
| Docker image size | 2.22 GB | 1.27 GB | 43% reduction |
| Fargate cold start | ~120s (0.25 vCPU) | ~40-57s (1 vCPU) | ~80s |
| Lambda cold start | 56s | 1.35s | ~54s |
make cold-start # Full measurement (stops tasks + cleans TaskState first)
make cold-start-warm # Warm start measurement (skip idle wait)
make task-status # Verify task state before measuring
make task-stop # Stop all tasks
make task-stop-recent # Stop only most recent task
Important: Always use make cold-start — never run the measurement script directly without stopping existing tasks first.
| Optimization | Savings | Technique |
|---|---|---|
| Remove AWS CLI | -358 MB | Replaced aws s3 sync with @aws-sdk/client-s3 Node.js code (s3-sync.ts) |
| chown layer optimization | -134 MB | COPY --chown=openclaw:openclaw instead of separate RUN chown |
| Pre-generated openclaw config | — | Handled in Dockerfile (skip runtime onboard) |
| start-openclaw.sh simplification | — | Removed S3 restore/onboard/sleep 2 |
| zstd compression (P6) | -16% (258→217 MB) | --compression zstd in docker buildx |
./scripts/deploy-image.sh --soci # Linux only, ~50% additional pull-time reduction
SOCI CI: .github/workflows/deploy-image.yml
| CPU | Memory | Cold Start |
|---|---|---|
| 0.25 vCPU | 512 MB | OOM Kill |
| 0.25 vCPU | 2048 MB | ~120s |
| 0.5 vCPU | 2048 MB | ~80s |
| 1 vCPU | 2048 MB | ~40-57s (default) |
| 2 vCPU | 4096 MB | Configurable via env |
Configure via env vars: FARGATE_CPU, FARGATE_MEMORY
Promise.all (~3-5s saved)Root cause: Bedrock auto-discovery scanned all AWS regions on startup → 56s delay
Fix: bedrockDiscovery.enabled: false in OpenClaw config → 1.35s cold start
{
"bedrockDiscovery": { "enabled": false }
}
runEmbeddedPiAgent() — bypasses Gateway WS server (no :18789 socket bind)extensionAPI.js via file:// URL (not in exports map)| Version | Cold Start | Status |
|---|---|---|
| v2026.2.9 | 65.3s | Working |
| v2026.2.13 | 57.9s | Working — pinned |
| v2026.2.14 | N/A | BROKEN (scope system) |
Pin location: packages/container/Dockerfile
ARG OPENCLAW_VERSION=2026.2.13
Configurable via build-arg or env var.
now < prewarmUntil