docker: unify per-service Dockerfiles into one multi-stage file

Collapse Dockerfile.web / Dockerfile.collab / Dockerfile.mcp into a single
docker/Dockerfile with shared `deps` stage and `web` / `collab` / `mcp`
runtime targets. BuildKit hashes the deps stage identically for all three
targets, so `pnpm install` runs once instead of three times in parallel —
fixes the OOM kill on the Coolify host (CT 107) during `docker compose build`.

Both docker-compose.yml (local dev) and docker-compose.coolify.yml now
select services via `target:` instead of separate dockerfile paths.

Made-with: Cursor
This commit is contained in:
Randall Stillwell 2026-04-27 14:38:39 -05:00
parent ba519a8b78
commit 9cdee7f84e
6 changed files with 186 additions and 170 deletions

168
docker/Dockerfile Normal file
View file

@ -0,0 +1,168 @@
# syntax=docker/dockerfile:1
#
# Single multi-stage Dockerfile for ECHODO. Build via Docker Compose with
# `target:` to select web | collab | mcp.
#
# Why one file instead of three?
# - The `deps` stage runs `pnpm install` ONCE for all 723 workspace packages.
# BuildKit shares it across every downstream target by content hash, so
# web + collab + mcp build in parallel without three concurrent installs
# blowing through the build host's RAM.
# - Per-app build stages (`web-build`, `collab-build`, `mcp-build`) only
# copy the source they actually need.
# - Final runtime stages (`web`, `collab`, `mcp`) pull just the built
# artifacts and the per-app node_modules they need at runtime.
ARG NODE_VERSION=20.18.1
# ============================================================================
# base — shared toolchain
# ============================================================================
FROM node:${NODE_VERSION}-alpine AS base
RUN corepack enable && corepack prepare pnpm@9.15.0 --activate
WORKDIR /app
# ============================================================================
# deps — shared pnpm install for the entire workspace
# Cached aggressively: invalidates only when package.json or lockfile changes.
# ============================================================================
FROM base AS deps
COPY package.json pnpm-lock.yaml pnpm-workspace.yaml turbo.json tsconfig.json ./
COPY apps/web/package.json ./apps/web/
COPY apps/collab-server/package.json ./apps/collab-server/
COPY apps/mcp-server/package.json ./apps/mcp-server/
COPY packages/database/package.json ./packages/database/
COPY packages/shared/package.json ./packages/shared/
COPY packages/ai/package.json ./packages/ai/
RUN pnpm install --frozen-lockfile
# ============================================================================
# web-build — Next.js production build → standalone output
# ============================================================================
FROM deps AS web-build
COPY apps/web ./apps/web
COPY packages/ai ./packages/ai
COPY packages/database ./packages/database
COPY packages/shared ./packages/shared
COPY docker/next.config.docker.ts ./apps/web/next.config.ts
ENV NEXT_TELEMETRY_DISABLED=1
ENV NODE_ENV=production
RUN pnpm exec turbo build --filter=@tasks/web
# ============================================================================
# collab-build — tsup build for Hocuspocus server
# ============================================================================
FROM deps AS collab-build
COPY apps/collab-server ./apps/collab-server
COPY packages/database ./packages/database
ENV NODE_ENV=production
RUN pnpm exec turbo build --filter=@tasks/collab-server
# ============================================================================
# mcp-build — tsup build for MCP server
# ============================================================================
FROM deps AS mcp-build
COPY apps/mcp-server ./apps/mcp-server
COPY packages/database ./packages/database
COPY packages/shared ./packages/shared
ENV NODE_ENV=production
RUN pnpm exec turbo build --filter=@tasks/mcp-server
# ============================================================================
# web — runtime image (Next.js standalone)
# ============================================================================
FROM node:${NODE_VERSION}-alpine AS web
RUN apk add --no-cache wget \
&& addgroup --system --gid 1001 nodejs \
&& adduser --system --uid 1001 nextjs
WORKDIR /app
ENV NODE_ENV=production
ENV PORT=3000
ENV HOSTNAME=0.0.0.0
COPY --from=web-build --chown=nextjs:nodejs /app/apps/web/.next/standalone ./
COPY --from=web-build --chown=nextjs:nodejs /app/apps/web/.next/static ./apps/web/.next/static
COPY --from=web-build --chown=nextjs:nodejs /app/apps/web/public ./apps/web/public
USER nextjs
EXPOSE 3000
HEALTHCHECK --interval=30s --timeout=5s --start-period=60s --retries=3 \
CMD wget -qO- --timeout=3 http://127.0.0.1:3000/ >/dev/null || exit 1
CMD ["node", "apps/web/server.js"]
# ============================================================================
# collab — runtime image (Hocuspocus / Yjs websocket)
# ============================================================================
FROM node:${NODE_VERSION}-alpine AS collab
RUN addgroup --system --gid 1001 nodejs \
&& adduser --system --uid 1001 collab --ingroup nodejs
WORKDIR /app
ENV NODE_ENV=production
ENV PORT=1234
COPY --from=collab-build --chown=collab:nodejs /app/node_modules ./node_modules
COPY --from=collab-build --chown=collab:nodejs /app/packages/database ./packages/database
COPY --from=collab-build --chown=collab:nodejs /app/apps/collab-server/dist ./apps/collab-server/dist
COPY --from=collab-build --chown=collab:nodejs /app/apps/collab-server/package.json ./apps/collab-server/package.json
COPY --from=collab-build --chown=collab:nodejs /app/apps/collab-server/node_modules ./apps/collab-server/node_modules
USER collab
EXPOSE 1234
HEALTHCHECK --interval=30s --timeout=5s --start-period=20s --retries=3 \
CMD node -e "const p=+(process.env.PORT||1234);require('net').createConnection(p,'127.0.0.1').on('connect',()=>process.exit(0)).on('error',()=>process.exit(1))"
WORKDIR /app/apps/collab-server
CMD ["node", "dist/index.mjs"]
# ============================================================================
# mcp — runtime image (Model Context Protocol server)
# ============================================================================
FROM node:${NODE_VERSION}-alpine AS mcp
ARG MCP_SERVER_PORT=3001
ENV MCP_SERVER_PORT=${MCP_SERVER_PORT}
RUN apk add --no-cache procps \
&& addgroup --system --gid 1001 nodejs \
&& adduser --system --uid 1001 mcp --ingroup nodejs
WORKDIR /app
ENV NODE_ENV=production
COPY --from=mcp-build --chown=mcp:nodejs /app/node_modules ./node_modules
COPY --from=mcp-build --chown=mcp:nodejs /app/packages/database ./packages/database
COPY --from=mcp-build --chown=mcp:nodejs /app/packages/shared ./packages/shared
COPY --from=mcp-build --chown=mcp:nodejs /app/apps/mcp-server/dist ./apps/mcp-server/dist
COPY --from=mcp-build --chown=mcp:nodejs /app/apps/mcp-server/package.json ./apps/mcp-server/package.json
COPY --from=mcp-build --chown=mcp:nodejs /app/apps/mcp-server/node_modules ./apps/mcp-server/node_modules
USER mcp
WORKDIR /app/apps/mcp-server
EXPOSE 3001
HEALTHCHECK --interval=30s --timeout=5s --start-period=20s --retries=3 \
CMD pgrep -f "node.*dist/index\\.js" > /dev/null || exit 1
CMD ["node", "dist/index.js"]

View file

@ -1,51 +0,0 @@
# syntax=docker/dockerfile:1
ARG NODE_VERSION=20.18.1
FROM node:${NODE_VERSION}-alpine AS base
RUN corepack enable && corepack prepare pnpm@9.15.0 --activate
WORKDIR /app
# --- Builder ---
# Install deps + build in one stage so pnpm's per-package node_modules
# (containing .bin symlinks like `tsup`) are intact.
FROM base AS builder
COPY package.json pnpm-lock.yaml pnpm-workspace.yaml turbo.json tsconfig.json ./
COPY apps/web/package.json ./apps/web/
COPY apps/collab-server/package.json ./apps/collab-server/
COPY apps/mcp-server/package.json ./apps/mcp-server/
COPY packages/database/package.json ./packages/database/
COPY packages/shared/package.json ./packages/shared/
COPY packages/ai/package.json ./packages/ai/
RUN pnpm install --frozen-lockfile
COPY apps/collab-server ./apps/collab-server
COPY packages/database ./packages/database
ENV NODE_ENV=production
RUN pnpm exec turbo build --filter=@tasks/collab-server
# --- Production runner ---
FROM node:${NODE_VERSION}-alpine AS runner
RUN addgroup --system --gid 1001 nodejs \
&& adduser --system --uid 1001 collab --ingroup nodejs
WORKDIR /app
ENV NODE_ENV=production
ENV PORT=1234
COPY --from=builder --chown=collab:nodejs /app/node_modules ./node_modules
COPY --from=builder --chown=collab:nodejs /app/packages/database ./packages/database
COPY --from=builder --chown=collab:nodejs /app/apps/collab-server/dist ./apps/collab-server/dist
COPY --from=builder --chown=collab:nodejs /app/apps/collab-server/package.json ./apps/collab-server/package.json
COPY --from=builder --chown=collab:nodejs /app/apps/collab-server/node_modules ./apps/collab-server/node_modules
USER collab
EXPOSE 1234
HEALTHCHECK --interval=30s --timeout=5s --start-period=20s --retries=3 \
CMD node -e "const p=+(process.env.PORT||1234);require('net').createConnection(p,'127.0.0.1').on('connect',()=>process.exit(0)).on('error',()=>process.exit(1))"
WORKDIR /app/apps/collab-server
CMD ["node", "dist/index.mjs"]

View file

@ -1,57 +0,0 @@
# syntax=docker/dockerfile:1
ARG NODE_VERSION=20.18.1
FROM node:${NODE_VERSION}-alpine AS base
RUN corepack enable && corepack prepare pnpm@9.15.0 --activate
WORKDIR /app
# --- Builder ---
# Install deps + build in one stage so pnpm's per-package node_modules
# (containing .bin symlinks like `tsup`) are intact.
FROM base AS builder
COPY package.json pnpm-lock.yaml pnpm-workspace.yaml turbo.json tsconfig.json ./
COPY apps/web/package.json ./apps/web/
COPY apps/collab-server/package.json ./apps/collab-server/
COPY apps/mcp-server/package.json ./apps/mcp-server/
COPY packages/database/package.json ./packages/database/
COPY packages/shared/package.json ./packages/shared/
COPY packages/ai/package.json ./packages/ai/
RUN pnpm install --frozen-lockfile
COPY apps/mcp-server ./apps/mcp-server
COPY packages/database ./packages/database
COPY packages/shared ./packages/shared
ENV NODE_ENV=production
RUN pnpm exec turbo build --filter=@tasks/mcp-server
# --- Production runner ---
FROM node:${NODE_VERSION}-alpine AS runner
ARG MCP_SERVER_PORT=3001
ENV MCP_SERVER_PORT=${MCP_SERVER_PORT}
RUN apk add --no-cache procps \
&& addgroup --system --gid 1001 nodejs \
&& adduser --system --uid 1001 mcp --ingroup nodejs
WORKDIR /app
ENV NODE_ENV=production
COPY --from=builder --chown=mcp:nodejs /app/node_modules ./node_modules
COPY --from=builder --chown=mcp:nodejs /app/packages/database ./packages/database
COPY --from=builder --chown=mcp:nodejs /app/packages/shared ./packages/shared
COPY --from=builder --chown=mcp:nodejs /app/apps/mcp-server/dist ./apps/mcp-server/dist
COPY --from=builder --chown=mcp:nodejs /app/apps/mcp-server/package.json ./apps/mcp-server/package.json
COPY --from=builder --chown=mcp:nodejs /app/apps/mcp-server/node_modules ./apps/mcp-server/node_modules
USER mcp
WORKDIR /app/apps/mcp-server
EXPOSE 3001
HEALTHCHECK --interval=30s --timeout=5s --start-period=20s --retries=3 \
CMD pgrep -f "node.*dist/index\\.js" > /dev/null || exit 1
CMD ["node", "dist/index.js"]

View file

@ -1,56 +0,0 @@
# syntax=docker/dockerfile:1
ARG NODE_VERSION=20.18.1
FROM node:${NODE_VERSION}-alpine AS base
RUN corepack enable && corepack prepare pnpm@9.15.0 --activate
WORKDIR /app
# --- Builder ---
# Install deps + build in one stage so pnpm's per-package node_modules
# (containing .bin symlinks like `next`, `tsup`) are intact.
# pnpm's content-addressable store keeps re-installs fast on cache hits.
FROM base AS builder
COPY package.json pnpm-lock.yaml pnpm-workspace.yaml turbo.json tsconfig.json ./
COPY apps/web/package.json ./apps/web/
COPY apps/collab-server/package.json ./apps/collab-server/
COPY apps/mcp-server/package.json ./apps/mcp-server/
COPY packages/database/package.json ./packages/database/
COPY packages/shared/package.json ./packages/shared/
COPY packages/ai/package.json ./packages/ai/
RUN pnpm install --frozen-lockfile
COPY apps/web ./apps/web
COPY packages/ai ./packages/ai
COPY packages/database ./packages/database
COPY packages/shared ./packages/shared
COPY docker/next.config.docker.ts ./apps/web/next.config.ts
ENV NEXT_TELEMETRY_DISABLED=1
ENV NODE_ENV=production
RUN pnpm exec turbo build --filter=@tasks/web
# --- Production runner ---
FROM node:${NODE_VERSION}-alpine AS runner
RUN apk add --no-cache wget \
&& addgroup --system --gid 1001 nodejs \
&& adduser --system --uid 1001 nextjs
WORKDIR /app
ENV NODE_ENV=production
ENV PORT=3000
ENV HOSTNAME=0.0.0.0
COPY --from=builder --chown=nextjs:nodejs /app/apps/web/.next/standalone ./
COPY --from=builder --chown=nextjs:nodejs /app/apps/web/.next/static ./apps/web/.next/static
COPY --from=builder --chown=nextjs:nodejs /app/apps/web/public ./apps/web/public
USER nextjs
EXPOSE 3000
HEALTHCHECK --interval=30s --timeout=5s --start-period=60s --retries=3 \
CMD wget -qO- --timeout=3 http://127.0.0.1:3000/ >/dev/null || exit 1
CMD ["node", "apps/web/server.js"]

View file

@ -16,12 +16,19 @@
# - No bundled postgres / redis (uses shared CT 102 services) # - No bundled postgres / redis (uses shared CT 102 services)
# - No host port mappings (Coolify's Traefik routes by container labels) # - No host port mappings (Coolify's Traefik routes by container labels)
# - Uses Coolify SERVICE_FQDN_* magic env vars so Coolify auto-wires Traefik labels # - Uses Coolify SERVICE_FQDN_* magic env vars so Coolify auto-wires Traefik labels
#
# Build strategy:
# All three services share docker/Dockerfile and select runtime via `target:`.
# The shared `deps` stage (pnpm install for the whole workspace) is hashed
# identically across targets, so BuildKit runs it ONCE even when web/collab/mcp
# build in parallel — keeps build-host RAM bounded.
services: services:
web: web:
build: build:
context: . context: .
dockerfile: docker/Dockerfile.web dockerfile: docker/Dockerfile
target: web
restart: unless-stopped restart: unless-stopped
environment: environment:
# Coolify FQDN magic var — set the domain in the Coolify UI for this service. # Coolify FQDN magic var — set the domain in the Coolify UI for this service.
@ -69,7 +76,8 @@ services:
collab: collab:
build: build:
context: . context: .
dockerfile: docker/Dockerfile.collab dockerfile: docker/Dockerfile
target: collab
restart: unless-stopped restart: unless-stopped
environment: environment:
# Separate FQDN for the websocket service, e.g. collab.echodo.stillwell.cloud # Separate FQDN for the websocket service, e.g. collab.echodo.stillwell.cloud
@ -96,7 +104,8 @@ services:
mcp: mcp:
build: build:
context: . context: .
dockerfile: docker/Dockerfile.mcp dockerfile: docker/Dockerfile
target: mcp
args: args:
MCP_SERVER_PORT: ${MCP_SERVER_PORT:-3001} MCP_SERVER_PORT: ${MCP_SERVER_PORT:-3001}
restart: unless-stopped restart: unless-stopped

View file

@ -47,7 +47,8 @@ services:
web: web:
build: build:
context: .. context: ..
dockerfile: docker/Dockerfile.web dockerfile: docker/Dockerfile
target: web
restart: unless-stopped restart: unless-stopped
environment: environment:
DATABASE_URL: ${DATABASE_URL} DATABASE_URL: ${DATABASE_URL}
@ -79,7 +80,8 @@ services:
collab: collab:
build: build:
context: .. context: ..
dockerfile: docker/Dockerfile.collab dockerfile: docker/Dockerfile
target: collab
restart: unless-stopped restart: unless-stopped
environment: environment:
DATABASE_URL: ${DATABASE_URL} DATABASE_URL: ${DATABASE_URL}
@ -110,7 +112,8 @@ services:
mcp: mcp:
build: build:
context: .. context: ..
dockerfile: docker/Dockerfile.mcp dockerfile: docker/Dockerfile
target: mcp
args: args:
MCP_SERVER_PORT: ${MCP_SERVER_PORT:-3001} MCP_SERVER_PORT: ${MCP_SERVER_PORT:-3001}
restart: unless-stopped restart: unless-stopped