Skip to main content

Building a Docker image

This section describes how to build a Backstage App into a deployable Docker image. It is split into three sections, first covering the host build approach, which is recommended due to its speed and more efficient and often simpler caching. The second section covers a full multi-stage Docker build, and the last section covers how to deploy the frontend and backend as separate images.

Something that goes for all of these docker deployment strategies is that they are stateless, so for a production deployment you will want to set up and connect to an external PostgreSQL instance where the backend plugins can store their state, rather than using SQLite.

This section assumes that an app has already been created with @backstage/create-app, in which the frontend is bundled and served from the backend. This is done using the @backstage/plugin-app-backend plugin, which also injects the frontend configuration into the app. This means that you only need to build and deploy a single container in a minimal setup of Backstage. If you wish to separate the serving of the frontend out from the backend, see the separate frontend topic below.

Host Build​

This section describes how to build a Docker image from a Backstage repo with most of the build happening outside of Docker. This is almost always the faster approach, as the build steps tend to execute faster, and it's possible to have more efficient caching of dependencies on the host, where a single change won't bust the entire cache.

The required steps in the host build are to install dependencies with yarn install, generate type definitions using yarn tsc, and build the backend package with yarn build:backend.

In a CI workflow it might look something like this, from the root:

yarn install --immutable

# tsc outputs type definitions to dist-types/ in the repo root, which are then consumed by the build
yarn tsc

# Build the backend, which bundles it all up into the packages/backend/dist folder.
# The configuration files here should match the one you use inside the Dockerfile below.
yarn build:backend --config ../../app-config.yaml --config ../../app-config.production.yaml

Once the host build is complete, we are ready to build our image. The following Dockerfile is included when creating a new app with @backstage/create-app:

FROM node:20-bookworm-slim

# Set Python interpreter for `node-gyp` to use
ENV PYTHON=/usr/bin/python3

# Install isolate-vm dependencies, these are needed by the @backstage/plugin-scaffolder-backend.
RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
    --mount=type=cache,target=/var/lib/apt,sharing=locked \
    apt-get update && \
    apt-get install -y --no-install-recommends python3 g++ build-essential && \
    rm -rf /var/lib/apt/lists/*

# Install sqlite3 dependencies. You can skip this if you don't use sqlite3 in the image,
# in which case you should also move better-sqlite3 to "devDependencies" in package.json.
RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
    --mount=type=cache,target=/var/lib/apt,sharing=locked \
    apt-get update && \
    apt-get install -y --no-install-recommends libsqlite3-dev && \
    rm -rf /var/lib/apt/lists/*

# From here on we use the least-privileged `node` user to run the backend.
USER node

# This should create the app dir as `node`.
# If it is instead created as `root` then the `tar` command below will fail: `can't create directory 'packages/': Permission denied`.
# If this occurs, then ensure BuildKit is enabled (`DOCKER_BUILDKIT=1`) so the app dir is correctly created as `node`.
WORKDIR /app

# Copy files needed by Yarn
COPY --chown=node:node .yarn ./.yarn
COPY --chown=node:node .yarnrc.yml ./

# This switches many Node.js dependencies to production mode.
ENV NODE_ENV=production

# This disables node snapshot for Node 20 to work with the Scaffolder
ENV NODE_OPTIONS="--no-node-snapshot"

# Copy repo skeleton first, to avoid unnecessary docker cache invalidation.
# The skeleton contains the package.json of each package in the monorepo,
# and along with yarn.lock and the root package.json, that's enough to run yarn install.
COPY --chown=node:node yarn.lock package.json packages/backend/dist/skeleton.tar.gz ./
RUN tar xzf skeleton.tar.gz && rm skeleton.tar.gz

RUN --mount=type=cache,target=/home/node/.cache/yarn,sharing=locked,uid=1000,gid=1000 \
    yarn workspaces focus --all --production && rm -rf "$(yarn cache clean)"

# This will include the examples, if you don't need these simply remove this line
COPY --chown=node:node examples ./examples

# Then copy the rest of the backend bundle, along with any other files we might want.
COPY --chown=node:node packages/backend/dist/bundle.tar.gz app-config*.yaml ./
RUN tar xzf bundle.tar.gz && rm bundle.tar.gz

CMD ["node", "packages/backend", "--config", "app-config.yaml", "--config", "app-config.production.yaml"]

For more details on how the backend:bundle command and the skeleton.tar.gz file works, see the backend:bundle command docs.

The Dockerfile is located at packages/backend/Dockerfile, but needs to be executed with the root of the repo as the build context, in order to get access to the root yarn.lock and package.json, along with any other files that might be needed, such as .npmrc.

The @backstage/create-app command adds the following .dockerignore in the root of the repo to speed up the build by reducing build context size:

.git
.yarn/cache
.yarn/install-state.gz
node_modules
packages/*/src
packages/*/node_modules
plugins
*.local.yaml

With the project built and the .dockerignore and Dockerfile in place, we are now ready to build the final image. From the root of the repo, execute the build:

docker image build . -f packages/backend/Dockerfile --tag backstage

To try out the image locally you can run the following:

docker run -it -p 7007:7007 backstage

You should then start to get logs in your terminal, and then you can open your browser at http://localhost:7007

Multi-stage Build​

Note

The .dockerignore is different in this setup, read on for more details.

This section describes how to set up a multi-stage Docker build that builds the entire project within Docker. This is typically slower than a host build, but is sometimes desired because Docker in Docker is not available in the build environment, or due to other requirements.

The build is split into three different stages, where the first stage finds all of the package.json files that are relevant for the initial install step enabling us to cache the initial yarn install that installs all dependencies. The second stage executes the build itself, and is similar to the steps we execute on the host in the host build. The third and final stage then packages it all together into the final image, and is similar to the Dockerfile of the host build.

The following Dockerfile executes the multi-stage build and should be added to the repo root:

# Stage 1 - Create yarn install skeleton layer
FROM node:20-bookworm-slim AS packages

WORKDIR /app
COPY package.json yarn.lock ./
COPY .yarn ./.yarn
COPY .yarnrc.yml ./

COPY packages packages

# Comment this out if you don't have any internal plugins
COPY plugins plugins

RUN find packages \! -name "package.json" -mindepth 2 -maxdepth 2 -exec rm -rf {} \+

# Stage 2 - Install dependencies and build packages
FROM node:20-bookworm-slim AS build

# Set Python interpreter for `node-gyp` to use
ENV PYTHON=/usr/bin/python3

# Install isolate-vm dependencies, these are needed by the @backstage/plugin-scaffolder-backend.
RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
    --mount=type=cache,target=/var/lib/apt,sharing=locked \
    apt-get update && \
    apt-get install -y --no-install-recommends python3 g++ build-essential && \
    rm -rf /var/lib/apt/lists/*

# Install sqlite3 dependencies. You can skip this if you don't use sqlite3 in the image,
# in which case you should also move better-sqlite3 to "devDependencies" in package.json.
RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
    --mount=type=cache,target=/var/lib/apt,sharing=locked \
    apt-get update && \
    apt-get install -y --no-install-recommends libsqlite3-dev && \
    rm -rf /var/lib/apt/lists/*

USER node
WORKDIR /app

COPY --from=packages --chown=node:node /app .
COPY --from=packages --chown=node:node /app/.yarn ./.yarn
COPY --from=packages --chown=node:node /app/.yarnrc.yml  ./

RUN --mount=type=cache,target=/home/node/.cache/yarn,sharing=locked,uid=1000,gid=1000 \
    yarn install --immutable

COPY --chown=node:node . .

RUN yarn tsc
RUN yarn --cwd packages/backend build

RUN mkdir packages/backend/dist/skeleton packages/backend/dist/bundle \
    && tar xzf packages/backend/dist/skeleton.tar.gz -C packages/backend/dist/skeleton \
    && tar xzf packages/backend/dist/bundle.tar.gz -C packages/backend/dist/bundle

# Stage 3 - Build the actual backend image and install production dependencies
FROM node:20-bookworm-slim

# Set Python interpreter for `node-gyp` to use
ENV PYTHON=/usr/bin/python3

# Install isolate-vm dependencies, these are needed by the @backstage/plugin-scaffolder-backend.
RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
    --mount=type=cache,target=/var/lib/apt,sharing=locked \
    apt-get update && \
    apt-get install -y --no-install-recommends python3 g++ build-essential && \
    rm -rf /var/lib/apt/lists/*

# Install sqlite3 dependencies. You can skip this if you don't use sqlite3 in the image,
# in which case you should also move better-sqlite3 to "devDependencies" in package.json.
RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
    --mount=type=cache,target=/var/lib/apt,sharing=locked \
    apt-get update && \
    apt-get install -y --no-install-recommends libsqlite3-dev && \
    rm -rf /var/lib/apt/lists/*

# From here on we use the least-privileged `node` user to run the backend.
USER node

# This should create the app dir as `node`.
# If it is instead created as `root` then the `tar` command below will
# fail: `can't create directory 'packages/': Permission denied`.
# If this occurs, then ensure BuildKit is enabled (`DOCKER_BUILDKIT=1`)
# so the app dir is correctly created as `node`.
WORKDIR /app

# Copy the install dependencies from the build stage and context
COPY --from=build --chown=node:node /app/.yarn ./.yarn
COPY --from=build --chown=node:node /app/.yarnrc.yml  ./
COPY --from=build --chown=node:node /app/yarn.lock /app/package.json /app/packages/backend/dist/skeleton/ ./
# Note: The skeleton bundle only includes package.json files -- if your app has
# plugins that define a `bin` export, the bin files need to be copied as well to
# be linked in node_modules/.bin during yarn install.

RUN --mount=type=cache,target=/home/node/.cache/yarn,sharing=locked,uid=1000,gid=1000 \
    yarn workspaces focus --all --production && rm -rf "$(yarn cache clean)"

# Copy the built packages from the build stage
COPY --from=build --chown=node:node /app/packages/backend/dist/bundle/ ./

# Copy any other files that we need at runtime
COPY --chown=node:node app-config*.yaml ./

# This will include the examples, if you don't need these simply remove this line
COPY --chown=node:node examples ./examples

# This switches many Node.js dependencies to production mode.
ENV NODE_ENV=production

# This disables node snapshot for Node 20 to work with the Scaffolder
ENV NODE_OPTIONS="--no-node-snapshot"

CMD ["node", "packages/backend", "--config", "app-config.yaml", "--config", "app-config.production.yaml"]

Note that a newly created Backstage app will typically not have a plugins/ folder, so you will want to comment that line out. This build also does not work in the main repo, since the backstage-cli which is used for the build doesn't end up being properly installed.

To speed up the build when not running in a fresh clone of the repo you should set up a .dockerignore. This one is different than the host build one, because we want to have access to the source code of all packages for the build. We can however ignore any existing build output or dependencies on the host. For our new .dockerignore, replace the contents of your existing one with this:

dist-types
node_modules
packages/*/dist
packages/*/node_modules
plugins/*/dist
plugins/*/node_modules
*.local.yaml

Once you have added both the Dockerfile and .dockerignore to the root of your project, run the following to build the container under a specified tag.

docker image build -t backstage .

To try out the image locally you can run the following:

docker run -it -p 7007:7007 backstage

You should then start to get logs in your terminal, and then you can open your browser at http://localhost:7007

Separate Frontend​

Note

This is an optional step, and you will lose out on the features of the @backstage/plugin-app-backend plugin. Most notably the frontend configuration will no longer be injected by the backend, you will instead need to use the correct configuration when building the frontend bundle.

It is sometimes desirable to serve the frontend separately from the backend, either from a separate image or for example a static file serving provider. The first step in doing so is to remove the app-backend plugin from the backend package, which is done as follows:

  1. Delete packages/backend/src/plugins/app.ts

  2. Remove the following line from packages/backend/src/index.ts:

    backend.add(import('@backstage/plugin-app-backend'));

  3. Remove the @backstage/plugin-app-backend and the app package dependency (e.g. app) from packages/backend/package.json. If you don't remove the app package dependency the app will still be built and bundled with the backend.

Once the app-backend is removed from the backend, you can use your favorite static file serving method for serving the frontend. An example of how to set up an NGINX image is available in the contrib folder in the main repo

Note that if you're building a separate docker build of the frontend you probably need to adjust .dockerignore appropriately. Most likely by making sure packages/app/dist is not ignored.

Troubleshooting Tips​

When building Docker images you may run into problems from time to time, there are two handy flags you can use to help:

  • --progress=plain: this will give you a more verbose output and not fold the logs into sections. This is very useful when have an error but it just shows you the last command and possibly an exit code. Using this flag you are more likely to see where the error actually is.
  • --no-cache: this will rebuild all the layers every time. This is helpful when you want to be sure that it's building from scratch.

Here's an example of these flags in use:

docker image build . -f packages/backend/Dockerfile --tag backstage --progress=plain --no-cache

Community Contributed Dockerfile Alternatives​

The Dockerfile mentioned above located in packages/backend is maintained by the maintainers of Backstage, however there are also community contributed Dockerfile alternatives located in contrib/docker. The Dockerfiles in contrib/docker are not maintained by the maintainers of Backstage and are not necessarily updated when the Dockerfile located in packages/backend is updated.

Minimal Hardened Image​

A contributed Dockerfile exists within the directory of contrib/docker/minimal-hardened-image which uses the wolfi-base image to reduce vulnerabilities. When this was contributed, this alternative Dockerfile reduced 98.2% of vulnerabilities in the built Backstage docker image when compared with the image built from packages/backend/Dockerfile.

To reduce maintenance, the digest of the image has been removed from the contrib/docker/minimal-hardened-image/Dockerfile file. A complete example with the digest would be cgr.dev/chainguard/wolfi-base:latest@sha256:3d6dece13cdb5546cd03b20e14f9af354bc1a56ab5a7b47dca3e6c1557211fcf and it is suggested to update the FROM line in the Dockerfile to use a digest. Please do a docker pull on the image to get the latest digest. Using the digest allows tools such as Dependabot or Renovate to know exactly which image digest is being utilized and allows for Pull Requests to be triggered when a new digest is available.

It is suggested to setup Dependabot/Renovate or a similar tool to ensure the image is kept up to date so that vulnerability fixes that have been addressed are pulled in frequently.