deer-flow/docker/docker-compose-dev.yaml
Nan Gao 063383d995
fix(deploy): wire the multi-replica prerequisites into the Helm chart and compose (#6347)
* fix(deploy): wire the multi-replica prerequisites into the Helm chart and compose

The chart shipped gateway.replicas: 1 with a README citing the long-closed
issue #3948, while its rendered config lacked run_ownership heartbeats and
db run events: raising replicas made every starting Pod write its peers'
lease-less runs off as orphans, and the Gateway's startup gate could not
notice because its worker count only sees one Pod.

Enable heartbeats and db run events in the default config, export
DEER_FLOW_MULTI_INSTANCE from the replica count for the startup gate,
generate and inject a shared AUTH_JWT_SECRET (per-Pod .jwt_secret files
logged users out across replicas), declare a surge-then-drain rollout
strategy, render a PodDisruptionBudget when replicas > 1, and bound
uvicorn's graceful shutdown inside a 90s termination grace period.
docker-compose gets the same uvicorn bound and a 60s stop_grace_period so
the 30s memory flush is no longer SIGKILLed. Docs and comments now list
what actually remains single-instance: IM channels, WeChat QR login and
browser tools.

* fix(deploy): honor existingAppSecret, declare multi-instance explicitly, cover the dev stack's shutdown

Follow-up to the multi-replica prerequisites wiring, addressing review:

- `deer-flow.appSecret` returns `existingAppSecret` when it is set. The
  gateway and frontend kept referencing the generated `<release>-app`
  Secret, which the chart stopped rendering, so every Pod sat in
  CreateContainerConfigError. The value is now declared in values.yaml.
- `gateway.multiInstance` (boolean, default null) overrides the
  replica-derived DEER_FLOW_MULTI_INSTANCE, PodDisruptionBudget and
  AUTH_JWT_SECRET requirement for deployments scaled with kubectl or an
  HPA, which change the replica count without a Helm upgrade.
- AUTH_JWT_SECRET is required, not optional, whenever the gateway is
  multi-instance: concurrently booting Pods otherwise race to write
  different `.jwt_secret` files on the home volume and sign sessions with
  different keys.
- `gateway.uvicornGracefulShutdownSeconds: 0` is honored instead of being
  swallowed by `default 10`.
- docker-compose-dev.yaml and dev-entrypoint.sh get the same uvicorn
  graceful-shutdown bound and `stop_grace_period: 60s`.
- The chart README documents the one-time sign-out when upgrading a
  release that predates AUTH_JWT_SECRET and how to seed the previous key
  into the Secret first.
- Tests read the shutdown budget from the Gateway through a shared helper
  instead of hard-coding 5/5/30, the raw-template string matches give way
  to `helm template` renders, and new renders cover multiInstance,
  existingAppSecret, a disabled PDB and a zero graceful-shutdown bound.

* docs(changelog): link the multi-replica deploy entry to #6347

* fix(deploy): budget every bounded teardown hook and flag the multiInstance PDB window

Review follow-ups on #6347:

- The shutdown budget modeled only channel stop + run drain + memory flush
  (40s). The lifespan also runs the startup trash sweep, the notification
  delivery worker stop, browser-session close and the MCP session-pool
  close, each bounded at _SHUTDOWN_HOOK_TIMEOUT_SECONDS, plus a 1s
  retrieval-warm wait: about 61s worst case. _gateway_shutdown_budget.py
  now reads all of them from the Gateway, a drift guard pins the hook
  count against app.py, and both compose files move stop_grace_period
  from 60s to 90s (Helm's 90s already had room: 5 + 10 + 61). Comments
  and the chart README cite the full arithmetic.
- gateway.multiInstance: true renders the PodDisruptionBudget immediately,
  which blocks drains of the single Pod until the scale-up happens; the
  README, values comment and both deployment guides say so and name
  podDisruptionBudget.enabled: false as the interim escape hatch.
2026-10-06 11:21:51 +08:00

277 lines
13 KiB
YAML

# DeerFlow Development Environment
#
# Requires Docker Compose >= 2.24. The env_file entries below use the long-form
# path/required syntax so a missing ../.env or ../frontend/.env does not abort
# Compose on Windows ("file not found" / "Le fichier spécifique est
# introuvable"). Clients older than 2.24 do not understand that field and
# reject this file with:
# validating docker-compose-dev.yaml: services.gateway.env_file.0 must be a string
# make docker-start checks the version and says so; direct `docker compose`
# callers get the message above instead.
#
# Supported entry: from the repository root, run `make docker-start`. That
# wrapper checks the Compose version, creates missing .env files, exports
# DEER_FLOW_ROOT, and invokes Compose from this directory with a relative
# filename.
#
# Running Compose yourself works too, but export DEER_FLOW_ROOT as the absolute
# checkout path first (see CONTRIBUTING.md for the Windows form). Unset, the
# DEER_FLOW_HOST_BASE_DIR and THREADS_HOST_PATH values below render as
# /backend/.deer-flow, so the provisioner bind-mounts the wrong host paths:
# DEER_FLOW_ROOT="$PWD" docker compose -f docker/docker-compose-dev.yaml up --build
# The -f path resolves against your shell's directory, not against this file:
# from the repository root use docker/docker-compose-dev.yaml; from inside
# docker/ use the bare docker-compose-dev.yaml, as make docker-start does.
#
# Services:
# - nginx: Reverse proxy (port 2026)
# - frontend: Frontend Next.js dev server (port 3000)
# - gateway: Backend Gateway API + agent runtime (port 8001)
# - redis: Redis Streams backend for cross-worker SSE stream bridge
# - provisioner (optional): Sandbox provisioner (creates Pods in host Kubernetes)
#
# Prerequisites:
# - Kubernetes cluster + kubeconfig are only required when using provisioner mode.
#
# Access: http://localhost:2026
services:
# ── Redis Stream Bridge ────────────────────────────────────────────────
redis:
image: redis:7-alpine
container_name: deer-flow-redis
command: ["redis-server", "--appendonly", "yes"]
volumes:
- redis-data:/data
healthcheck:
test: ["CMD", "redis-cli", "ping"]
interval: 5s
timeout: 3s
retries: 10
networks:
- deer-flow-dev
restart: unless-stopped
# ── Sandbox Provisioner ────────────────────────────────────────────────
# Manages per-sandbox Pod + Service lifecycle in the host Kubernetes
# cluster via the K8s API.
# Backend accesses sandboxes directly via host.docker.internal:{NodePort}.
provisioner:
build:
context: ./provisioner
dockerfile: Dockerfile
args:
APT_MIRROR: ${APT_MIRROR:-}
container_name: deer-flow-provisioner
volumes:
- ~/.kube/config:/root/.kube/config:ro
environment:
- K8S_NAMESPACE=deer-flow
- SANDBOX_IMAGE=enterprise-public-cn-beijing.cr.volces.com/vefaas-public/all-in-one-sandbox:latest
# Optional lark-cli init image (Pattern A). Empty ⇒ legacy runtime mount.
# Set to a published tag (e.g. deer-flow/lark-cli-init:v1.0.65) to provision
# the sandbox lark-cli runtime via an init container + emptyDir.
- LARK_CLI_INIT_IMAGE=${LARK_CLI_INIT_IMAGE:-}
# Optional lark-cli broker image (Pattern B, issue #4338). Empty ⇒ broker
# off. Supersedes LARK_CLI_INIT_IMAGE when both are set.
- LARK_CLI_BROKER_IMAGE=${LARK_CLI_BROKER_IMAGE:-}
# Host paths for K8s HostPath volumes (must be absolute paths accessible by K8s node)
# On Docker Desktop/OrbStack, use your actual host paths like /Users/username/...
# Set these in your shell before running docker-compose:
# export DEER_FLOW_ROOT=/absolute/path/to/deer-flow
# Deliberately left without a ${DEER_FLOW_ROOT:-...} fallback. The obvious
# candidate, $PWD, is exported by POSIX shells but not by PowerShell or
# cmd, so the default would resolve to an empty string on Windows — the
# platform this variable exists to get right. A wrong-but-plausible host
# path mounts an empty directory instead of failing, so we require the
# caller to be explicit; make docker-start sets it for you.
- THREADS_HOST_PATH=${DEER_FLOW_ROOT}/backend/.deer-flow/threads
# Per-user data base directory for user-scoped skill mounts
- DEER_FLOW_HOST_BASE_DIR=${DEER_FLOW_ROOT}/backend/.deer-flow
# Production: use PVC instead of hostPath to avoid data loss on node failure.
# When set, hostPath vars above are ignored for the corresponding volume.
# USERDATA_PVC_NAME uses subPath (deer-flow/users/{user_id}/threads/{thread_id}/user-data) automatically.
# - SKILLS_PVC_NAME=deer-flow-skills-pvc
# - USERDATA_PVC_NAME=deer-flow-userdata-pvc
- KUBECONFIG_PATH=/root/.kube/config
- NODE_HOST=host.docker.internal
# Override K8S API server URL since kubeconfig uses 127.0.0.1
# which is unreachable from inside the container
- K8S_API_SERVER=https://host.docker.internal:26443
# Optional: set PROVISIONER_API_KEY in .env to enable provisioner auth.
# The same value must be set on the gateway side via config.yaml sandbox.provisioner_api_key.
- PROVISIONER_API_KEY=${PROVISIONER_API_KEY:-}
env_file:
- path: ../.env
required: false
extra_hosts:
- "host.docker.internal:host-gateway"
networks:
- deer-flow-dev
restart: unless-stopped
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8002/health"]
interval: 10s
timeout: 5s
retries: 6
start_period: 15s
# ── Reverse Proxy ──────────────────────────────────────────────────────
# Routes API traffic to gateway and (optionally) provisioner.
nginx:
image: nginx:alpine
container_name: deer-flow-nginx
# Loopback-only by default; see the note in docker-compose.yaml. Override
# with BIND_HOST when you deliberately need the dev stack reachable from
# another machine.
ports:
- "${BIND_HOST:-127.0.0.1}:${PORT:-2026}:2026"
volumes:
- ./nginx/nginx.conf:/etc/nginx/nginx.conf.template:ro
command:
- sh
- -c
- |
set -e
cp /etc/nginx/nginx.conf.template /etc/nginx/nginx.conf
test -e /proc/net/if_inet6 || sed -i '/^[[:space:]]*listen[[:space:]]\+\[::\]:2026[[:space:]]/d' /etc/nginx/nginx.conf
exec nginx -g 'daemon off;'
depends_on:
- frontend
- gateway
networks:
- deer-flow-dev
restart: unless-stopped
# Frontend - Next.js Development Server
frontend:
build:
context: ../
dockerfile: frontend/Dockerfile
target: dev
args:
PNPM_STORE_PATH: ${PNPM_STORE_PATH:-/root/.local/share/pnpm/store}
NPM_REGISTRY: ${NPM_REGISTRY:-}
container_name: deer-flow-frontend
command: sh -c "cd frontend && pnpm run dev > /app/logs/frontend.log 2>&1"
volumes:
- ../frontend/src:/app/frontend/src
- ../frontend/public:/app/frontend/public
- ../frontend/next.config.js:/app/frontend/next.config.js:ro
- ../logs:/app/logs
# Mount pnpm store for caching
- ${PNPM_STORE_PATH:-~/.local/share/pnpm/store}:/root/.local/share/pnpm/store
working_dir: /app
environment:
- NODE_ENV=development
- WATCHPACK_POLLING=true
- CI=true
- DEER_FLOW_INTERNAL_GATEWAY_BASE_URL=http://gateway:8001
- DEER_FLOW_DEV_ALLOWED_ORIGINS=${DEER_FLOW_DEV_ALLOWED_ORIGINS:-127.0.0.1,::1}
env_file:
- path: ../frontend/.env
required: false
networks:
- deer-flow-dev
restart: unless-stopped
# Backend - Gateway API
gateway:
build:
context: ../
dockerfile: backend/Dockerfile
target: dev
# cache_from disabled - requires manual setup: mkdir -p /tmp/docker-cache-gateway
args:
APT_MIRROR: ${APT_MIRROR:-}
UV_IMAGE: ${UV_IMAGE:-ghcr.io/astral-sh/uv:0.11.1}
UV_INDEX_URL: ${UV_INDEX_URL:-https://pypi.org/simple}
NPM_REGISTRY: ${NPM_REGISTRY:-}
LARK_CLI_NPM_VERSION: ${LARK_CLI_NPM_VERSION:-1.0.65}
container_name: deer-flow-gateway
# Startup logic lives in docker/dev-entrypoint.sh — UV_EXTRAS validation,
# `uv sync --all-packages`, .venv self-heal, and uvicorn handoff. Keeps
# this file readable and lets the script be linted (shellcheck-clean).
# See PR #2767 / Issue #2754.
command: ["sh", "/usr/local/bin/dev-entrypoint.sh"]
# Docker's 10s default SIGKILLed the 30s memory queue flush
# (memory.shutdown_flush_timeout_seconds) on every `make docker-restart`;
# dev-entrypoint.sh bounds uvicorn's graceful shutdown to 10s inside this.
# Must cover that bound plus the lifespan's worst case (~61s, see the
# gateway comment in docker-compose.yaml) with slack.
stop_grace_period: 90s
volumes:
# Mount the dev entrypoint as a read-only file so edits to the script
# take effect on `make docker-restart` without requiring an image rebuild.
- ./dev-entrypoint.sh:/usr/local/bin/dev-entrypoint.sh:ro
- ../backend/:/app/backend/
# Preserve the .venv built during Docker image build — mounting the full backend/
# directory above would otherwise shadow it with the (empty) host directory.
- gateway-venv:/app/backend/.venv
# Mount the project directory instead of its mutable config files individually.
# Host editors commonly replace files on save; a directory bind keeps those
# replacements visible inside Docker Desktop/WSL containers.
- ../:/app/project
- ../skills:/app/skills
- ../logs:/app/logs
# Use a Docker-managed uv cache volume instead of a host bind mount.
# On macOS/Docker Desktop, uv may fail to create symlinks inside shared
# host directories, which causes startup-time `uv sync` to crash.
- gateway-uv-cache:/root/.cache/uv
# DooD: the host Docker socket is NOT mounted by default. It is added only
# for aio (pure-DooD) sandbox mode via the opt-in docker-compose.dood.yaml
# overlay (appended by scripts/docker.sh). See backend/docs/CONFIGURATION.md
# CLI auth dirs (Claude Code / Codex) are NOT mounted by default: they
# expose the entire ~/.claude and ~/.codex (history, projects, global
# config, credentials) into the container. Mount them only when you use
# the Claude/Codex CLI login as a model provider or ACP agent, via the
# opt-in docker-compose.cli-auth.yaml overlay. Prefer an env token
# (CLAUDE_CODE_OAUTH_TOKEN, see .env.example / backend/docs/CONFIGURATION.md).
working_dir: /app
environment:
- CI=true
- DEER_FLOW_PROJECT_ROOT=/app
- DEER_FLOW_HOME=/app/backend/.deer-flow
- DEER_FLOW_CONFIG_PATH=/app/project/config.yaml
- DEER_FLOW_EXTENSIONS_CONFIG_PATH=/app/project/extensions_config.json
- DEER_FLOW_STREAM_BRIDGE_REDIS_URL=${DEER_FLOW_STREAM_BRIDGE_REDIS_URL:-redis://redis:6379/0}
- DEER_FLOW_CHANNELS_LANGGRAPH_URL=${DEER_FLOW_CHANNELS_LANGGRAPH_URL:-http://gateway:8001/api}
- DEER_FLOW_CHANNELS_GATEWAY_URL=${DEER_FLOW_CHANNELS_GATEWAY_URL:-http://gateway:8001}
- DEER_FLOW_INTERNAL_AUTH_TOKEN=${DEER_FLOW_INTERNAL_AUTH_TOKEN:-}
- DEER_FLOW_HOST_BASE_DIR=${DEER_FLOW_ROOT}/backend/.deer-flow
- DEER_FLOW_SANDBOX_HOST=host.docker.internal
# Pass PROVISIONER_API_KEY into the gateway container so config.yaml can reference it
# as sandbox.provisioner_api_key: $PROVISIONER_API_KEY
- PROVISIONER_API_KEY=${PROVISIONER_API_KEY:-}
# Proxy values (HTTP_PROXY/HTTPS_PROXY/ALL_PROXY) are inherited from ../.env via env_file.
# Only NO_PROXY is declared here so internal service hostnames are always exempt from the proxy.
- NO_PROXY=${NO_PROXY:-}${NO_PROXY:+,}localhost,127.0.0.1,::1,gateway,frontend,nginx,provisioner,openviking,host.docker.internal
- no_proxy=${no_proxy:-}${no_proxy:+,}localhost,127.0.0.1,::1,gateway,frontend,nginx,provisioner,openviking,host.docker.internal
env_file:
- path: ../.env
required: false
extra_hosts:
# For Linux: map host.docker.internal to host gateway
- "host.docker.internal:host-gateway"
depends_on:
redis:
condition: service_healthy
networks:
- deer-flow-dev
restart: unless-stopped
volumes:
# Persist .venv across container restarts so dependencies installed during
# image build are not shadowed by the host backend/ directory mount.
gateway-venv:
gateway-uv-cache:
redis-data:
networks:
deer-flow-dev:
driver: bridge
ipam:
config:
- subnet: 192.168.200.0/24