mirror of
https://github.com/OpenHands/OpenHands.git
synced 2026-10-07 16:19:05 +08:00
7dbe8fd7de
* fix: populate <RUNTIME_SERVICES> in static builds so automations work The agent's <RUNTIME_SERVICES> system-prompt block is built from VITE_RUNTIME_SERVICES_INFO, which the dev launchers set at build time. Static builds (the Docker image and the published binary) run `npm run build` without it, so the block is dropped and the agent does not know how to reach the local automation backend — it falls back to the cloud API (app.all-hands.dev) and automation creation fails. Inject the info at serve time, mirroring the existing --session-api-key path: - scripts/static-server.mjs: new --runtime-services-info flag injects the JSON as window.__AGENT_CANVAS_RUNTIME_SERVICES_INFO__ - src/api/agent-server-adapter.ts: parseRuntimeServicesInfo() falls back to that window global when the env var is empty - docker/entrypoint.sh: builds the JSON from the sandbox-facing service URLs (AGENT_SERVER_URL, AUTOMATION_BASE_URL) and passes it to the static server(s) Refs #1098 * refactor: extract runtime-services-info into a shared module Replace the inline JSON building in docker/entrypoint.sh with a single source of truth for the <RUNTIME_SERVICES> shape. buildRuntimeServicesInfo now lives in scripts/runtime-services-info.mjs (moved out of dev-safe.mjs, which re-exports it for back-compat) and gains: - optional full-URL overrides (agentServerUrl, automation.url) so the container can pass its runtime-resolved AGENT_SERVER_URL / AUTOMATION_BASE_URL and use 127.0.0.1 (avoiding IPv6 loopback) instead of build-time ports, and - a CLI entrypoint so docker/entrypoint.sh emits the JSON by running the same builder the dev stack uses, rather than a hand-rolled node -e blob. The Dockerfile ships the new dependency-free module into the image. * fix: pass --runtime-services-info in static mode and verify in automation e2e - startStaticFrontend() in dev-with-automation.mjs now builds the runtime-services info JSON via buildAutomationRuntimeServicesInfo() and passes it to static-server.mjs via --runtime-services-info. Without this, the npm binary / dev:static path served pre-built frontends that never populated the agent's <RUNTIME_SERVICES> system-prompt block (the Docker entrypoint already did this). - The mock-LLM automation e2e test now verifies that the <RUNTIME_SERVICES> block is present in the system messages sent to the LLM, and that it includes the expected service entries (Agent Server, Automation backend, /api/automation). Co-authored-by: openhands <openhands@all-hands.dev> * docs: update AGENTS.md with runtime-services-info plumbing changes Co-authored-by: openhands <openhands@all-hands.dev> --------- Co-authored-by: Rohit Malhotra <rohitvinodmalhotra@gmail.com> Co-authored-by: openhands <openhands@all-hands.dev>
143 lines
6.7 KiB
Docker
143 lines
6.7 KiB
Docker
# syntax=docker/dockerfile:1.7
|
|
|
|
# ═══════════════════════════════════════════════════════════════════════════════
|
|
# agent-canvas all-in-one Docker image
|
|
#
|
|
# Combines three services into a single image:
|
|
# 1. Agent Server — from ghcr.io/openhands/agent-server (upstream SDK image)
|
|
# 2. Automation — installed via pip from openhands-automation
|
|
# 3. Frontend — agent-canvas static build served by Node.js
|
|
#
|
|
# The entrypoint starts all three services and an ingress proxy that unifies
|
|
# them behind a single port (default 8000):
|
|
# /api/automation/* → automation backend (:18001)
|
|
# /api/*, /sockets → agent server (:18000)
|
|
# /* (default) → static frontend + SPA fallback
|
|
# ═══════════════════════════════════════════════════════════════════════════════
|
|
|
|
# ── Build args ────────────────────────────────────────────────────────────────
|
|
# No hardcoded defaults — values are derived from config/defaults.json.
|
|
# CI passes these via --build-arg; for local builds use:
|
|
# node scripts/docker-build.mjs (recommended, reads JSON for you)
|
|
# docker build --build-arg AGENT_SERVER_IMAGE=... --build-arg AUTOMATION_VERSION=... -f docker/Dockerfile .
|
|
ARG AGENT_SERVER_IMAGE
|
|
ARG AUTOMATION_VERSION
|
|
|
|
# ── Stage 1: Build frontend ──────────────────────────────────────────────────
|
|
FROM node:24-slim AS frontend-build
|
|
|
|
WORKDIR /build
|
|
|
|
# Cache-friendly: package files first
|
|
COPY package.json package-lock.json ./
|
|
RUN npm ci
|
|
|
|
# Copy everything needed for the build
|
|
COPY . .
|
|
|
|
# Build the static frontend.
|
|
# VITE_APP_ENV controls the PostHog telemetry key baked into the bundle:
|
|
# "production" → prod key (set by CI for tagged releases)
|
|
# anything else → staging key (default for PR / main / local builds)
|
|
ARG VITE_APP_ENV=""
|
|
ENV VITE_APP_ENV=${VITE_APP_ENV}
|
|
RUN npm run build
|
|
|
|
# ── Stage 1b: Generate shell-sourceable defaults from config/defaults.json ──
|
|
# This avoids needing jq/python at container runtime to parse the JSON.
|
|
FROM node:24-slim AS config-gen
|
|
COPY config/defaults.json /tmp/defaults.json
|
|
RUN node -e " \
|
|
const c = JSON.parse(require('fs').readFileSync('/tmp/defaults.json','utf-8')); \
|
|
const lines = [ \
|
|
'CONFIG_AGENT_SERVER_PORT=' + c.ports.agentServer, \
|
|
'CONFIG_AUTOMATION_PORT=' + c.ports.automation, \
|
|
'CONFIG_PROXY_PORT=' + c.ports.proxy, \
|
|
'CONFIG_STATE_SUBDIR=' + c.paths.stateSubdir, \
|
|
'CONFIG_CONVERSATIONS=' + c.paths.conversations, \
|
|
'CONFIG_BASH_EVENTS=' + c.paths.bashEvents, \
|
|
'CONFIG_AUTOMATION_DB=' + c.paths.automationDb, \
|
|
]; \
|
|
require('fs').writeFileSync('/tmp/defaults.env', lines.join('\n') + '\n'); \
|
|
"
|
|
|
|
# ── Stage 2: Combined image ──────────────────────────────────────────────────
|
|
FROM ${AGENT_SERVER_IMAGE} AS final
|
|
|
|
ARG AUTOMATION_VERSION
|
|
ARG OPENHANDS_BUILD_GIT_SHA=unknown
|
|
ARG OPENHANDS_BUILD_GIT_REF=unknown
|
|
|
|
ARG AGENT_CANVAS_VERSION=dev
|
|
LABEL org.opencontainers.image.title="agent-canvas"
|
|
LABEL org.opencontainers.image.description="All-in-one agent-canvas: Agent Server + Automation + Frontend"
|
|
LABEL org.opencontainers.image.source="https://github.com/OpenHands/agent-canvas"
|
|
LABEL org.opencontainers.image.version="${AGENT_CANVAS_VERSION}"
|
|
LABEL org.opencontainers.image.revision="${OPENHANDS_BUILD_GIT_SHA}"
|
|
|
|
ENV AGENT_CANVAS_BUILD_GIT_SHA=${OPENHANDS_BUILD_GIT_SHA}
|
|
ENV AGENT_CANVAS_BUILD_GIT_REF=${OPENHANDS_BUILD_GIT_REF}
|
|
|
|
USER root
|
|
|
|
# Install system deps required by automation's transitive dependencies
|
|
# (asyncpg needs libpq, which the agent-server base image may not include).
|
|
RUN if command -v apt-get >/dev/null 2>&1; then \
|
|
apt-get update && \
|
|
apt-get install -y --no-install-recommends libpq-dev && \
|
|
rm -rf /var/lib/apt/lists/*; \
|
|
fi
|
|
|
|
# Install automation server via pip.
|
|
# Shared deps (openhands-sdk, fastapi, uvicorn, pydantic, httpx, …) are
|
|
# already satisfied by the agent-server base image, so only automation-
|
|
# specific packages (asyncpg, sqlalchemy, boto3, gcloud, …) are added.
|
|
RUN uv pip install --system "openhands-automation==${AUTOMATION_VERSION}" 2>/dev/null \
|
|
|| pip install --no-cache-dir "openhands-automation==${AUTOMATION_VERSION}"
|
|
|
|
# Copy the frontend build output.
|
|
# react-router.config.ts unpacks build/client/ into build/ for non-Vercel builds.
|
|
COPY --from=frontend-build /build/build /opt/agent-canvas/frontend
|
|
|
|
# Copy the static-server script (serves frontend + proxies to backends)
|
|
COPY scripts/static-server.mjs /opt/agent-canvas/static-server.mjs
|
|
|
|
# Copy the runtime-services-info builder (entrypoint.sh runs it as a CLI to
|
|
# emit the agent's <RUNTIME_SERVICES> block; same builder the dev stack uses).
|
|
COPY scripts/runtime-services-info.mjs /opt/agent-canvas/runtime-services-info.mjs
|
|
|
|
# Copy custom tools (e.g. canvas_ui_tool.py) so the agent-server can import
|
|
# them via tool_module_qualnames. OH_EXTRA_PYTHON_PATH is set in entrypoint.sh.
|
|
COPY tools/ /opt/agent-canvas/tools/
|
|
|
|
# Copy generated defaults.env (from config/defaults.json via config-gen stage)
|
|
COPY --from=config-gen /tmp/defaults.env /opt/agent-canvas/defaults.env
|
|
|
|
# Copy the entrypoint
|
|
COPY docker/entrypoint.sh /opt/agent-canvas/entrypoint.sh
|
|
RUN chmod +x /opt/agent-canvas/entrypoint.sh
|
|
|
|
# Pre-create persistence directories with correct ownership so the
|
|
# openhands user can write to them even when Docker creates anonymous
|
|
# volumes (which default to root).
|
|
RUN mkdir -p /home/openhands/.openhands/agent-canvas/conversations \
|
|
/home/openhands/.openhands/agent-canvas/bash_events \
|
|
/home/openhands/.openhands/automation \
|
|
/projects && \
|
|
chown -R openhands:openhands /home/openhands/.openhands /projects
|
|
|
|
USER openhands
|
|
|
|
# Persistence volumes:
|
|
# /home/openhands/.openhands — settings, secrets, conversations, automation DB
|
|
# /projects — user code the agent can read/edit
|
|
# Bind-mount these for data to survive container restarts:
|
|
# docker run -v ~/.openhands:/home/openhands/.openhands -v ~/projects:/projects ...
|
|
VOLUME ["/home/openhands/.openhands", "/projects"]
|
|
|
|
# The entrypoint starts all services and the ingress proxy.
|
|
# Port 8000 is the unified entry point.
|
|
EXPOSE 8000
|
|
|
|
ENTRYPOINT ["tini", "--", "/opt/agent-canvas/entrypoint.sh"]
|