ci: build cloud images independently for each merge

Co-Authored-By: Paperclip <noreply@paperclip.ing>
This commit is contained in:
Devin Foley 2026-09-10 18:16:46 -07:00
parent 01c9dbcd83
commit c6050e2e27
6 changed files with 280 additions and 230 deletions

View File

@ -6,6 +6,7 @@ const workflows = [
'.github/workflows/refresh-lockfile.yml',
'.github/workflows/pr-trusted.yml',
'.github/workflows/docker.yml',
'.github/workflows/docker-cloud.yml',
];
test('lockfile repair workflows resolve dependencies instead of updating metadata only', async () => {

246
.github/workflows/docker-cloud.yml vendored Normal file
View File

@ -0,0 +1,246 @@
name: Docker cloud
on:
push:
branches: [master]
workflow_dispatch:
workflow_call:
permissions: {}
# Independent SHAs can build immediately on separate existing hosted runners.
# Repeated requests for the same source serialize without cancelling a build.
# No mutable canary channel is promoted here; docker.yml owns that operation.
concurrency:
group: docker-cloud-${{ github.sha }}
cancel-in-progress: false
jobs:
build-and-push-cloud:
runs-on: ubuntu-latest
timeout-minutes: 60
permissions:
contents: read
packages: write
steps:
- name: Checkout
uses: actions/checkout@v7
with:
# Full history and tags so `git describe` below can compute the
# release version to stamp into the image.
fetch-depth: 0
# `.git` is dockerignored, so a running image cannot derive its own
# version and otherwise reports the source package.json placeholder in
# analytics and the debug panel. Compute it here from the pristine
# checkout (real CalVer drift from the nearest release tag) and pass it
# into the build. Empty when no release tag is reachable — the server
# then keeps its existing fallbacks.
- name: Compute build version
id: build-version
run: |
set -euo pipefail
case "${GITHUB_REF}" in
refs/tags/nightly/v*)
# Lane tags carry the exact published version; stamp it verbatim
# instead of describing drift from the nearest stable tag.
version="${GITHUB_REF#refs/tags/nightly/v}"
;;
refs/tags/beta/v*)
version="${GITHUB_REF#refs/tags/beta/v}"
;;
*)
version="$(git describe --tags --match 'v*' --long --dirty 2>/dev/null || true)"
;;
esac
echo "version=${version}" >> "$GITHUB_OUTPUT"
echo "Stamping build version: ${version:-<none>}"
# ISO week stamp for the Dockerfile's tool layer: the layer caches
# across commits and re-pulls the @latest CLI tools when the week rolls
# over, instead of on every build.
- name: Compute tool cache epoch
id: tools-epoch
run: echo "epoch=$(date -u +%G-W%V)" >> "$GITHUB_OUTPUT"
- name: Setup pnpm
uses: pnpm/action-setup@v6
with:
version: 9.15.4
run_install: false
# No dependency cache here: this workflow publishes release images, and
# restoring a shared Actions cache into the build inputs would let a
# poisoned cache entry reach the published artifact.
- name: Setup Node.js
uses: actions/setup-node@v7
with:
node-version: 24
- name: Refresh lockfile for Docker build context
run: |
set -euo pipefail
pnpm install --resolution-only --ignore-scripts --no-frozen-lockfile
changed="$(git status --porcelain)"
if [ -z "$changed" ]; then
echo "Lockfile already matches package metadata."
exit 0
fi
if printf '%s\n' "$changed" | grep -Fvq ' pnpm-lock.yaml'; then
echo "Unexpected files changed during lockfile refresh:"
echo "$changed"
exit 1
fi
echo "Using refreshed pnpm-lock.yaml in the Docker build context."
- name: Free runner disk
run: |
set -euo pipefail
echo "Disk before cleanup:"
df -h
pnpm store prune || true
sudo apt-get clean || true
sudo rm -rf \
/usr/share/dotnet \
/usr/share/swift \
/usr/local/lib/android \
/usr/local/share/boost \
/usr/local/share/powershell \
/opt/ghc \
/opt/hostedtoolcache/CodeQL \
/opt/hostedtoolcache/PyPy \
/opt/hostedtoolcache/Ruby || true
docker system prune -af || true
echo "Disk after cleanup:"
df -h
- name: Login to GitHub Container Registry
uses: docker/login-action@v4
with:
registry: ghcr.io
username: ${{ github.repository_owner }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v4
# Deployment tooling reads these labels from the registry to verify an
# image's schema expectations against a migrator before deploying it,
# without pulling the image. The server refuses to start when the
# database is missing bundled migrations, so orchestrators need a cheap
# way to check image/migrator compatibility up front.
- name: Compute schema migration labels
id: schema
run: |
set -euo pipefail
last=$(ls packages/db/src/migrations/*.sql | sed 's|.*/||' | LC_ALL=C sort | tail -1)
count=$(ls packages/db/src/migrations/*.sql | wc -l | tr -d ' ')
echo "last=${last}" >> "$GITHUB_OUTPUT"
echo "count=${count}" >> "$GITHUB_OUTPUT"
# Published under the same lane tag set as the self-hosted image, with a
# `-cloud` suffix (nightly-cloud, latest-cloud, <version>-cloud,
# sha-<short>-cloud). `:canary-cloud` follows the same retag-step
# ownership rule as `:canary` above.
- name: Docker meta (cloud)
id: meta-cloud
uses: docker/metadata-action@v6
with:
images: ghcr.io/${{ github.repository }}
flavor: |
suffix=-cloud,onlatest=true
tags: |
type=raw,value=nightly,enable=${{ startsWith(github.ref, 'refs/tags/nightly/v') }}
type=raw,value=beta,enable=${{ startsWith(github.ref, 'refs/tags/beta/v') }}
type=raw,value=latest,enable=${{ startsWith(github.ref, 'refs/tags/v') }}
type=semver,pattern={{version}},enable=${{ startsWith(github.ref, 'refs/tags/v') }}
type=semver,pattern={{major}}.{{minor}},enable=${{ startsWith(github.ref, 'refs/tags/v') }}
type=sha
labels: |
io.github.paperclipai.schema.last-migration=${{ steps.schema.outputs.last }}
io.github.paperclipai.schema.migration-count=${{ steps.schema.outputs.count }}
- name: Build and push (cloud)
id: build-cloud
uses: docker/build-push-action@v7
with:
context: .
target: cloud
# Space-separated sandbox-provider directory names to build into
# the variant; add here when managed deployments need another.
# CLOUD_BUNDLED_SERVER_DEPS names the optional peer packages the
# variant installs from server/package.json's declared version;
# add another name there when a managed tenant needs it.
build-args: |
CLOUD_BUNDLED_PLUGINS=daytona
CLOUD_BUNDLED_SERVER_DEPS=@sentry/node
PAPERCLIP_BUILD_VERSION=${{ steps.build-version.outputs.version }}
PAPERCLIP_BUILD_COMMIT=${{ github.sha }}
CLI_TOOLS_CACHE_EPOCH=${{ steps.tools-epoch.outputs.epoch }}
# amd64 only, unlike the self-hosted image above: the cloud variant
# is consumed exclusively by managed-deployment hosts, which run
# amd64. The QEMU-emulated arm64 half dominated this job's wall
# clock, and dropping it roughly halves time-to-deployable-image.
platforms: linux/amd64
push: true
# Registry-backed BuildKit cache, separate ref from the self-hosted
# job so the two parallel builds never clobber each other's cache
# manifest (see the rationale on the job above).
cache-from: type=registry,ref=ghcr.io/${{ github.repository }}:buildcache-cloud
cache-to: type=registry,ref=ghcr.io/${{ github.repository }}:buildcache-cloud,mode=max
tags: ${{ steps.meta-cloud.outputs.tags }}
labels: ${{ steps.meta-cloud.outputs.labels }}
# The cloud target installs @sentry/node at the version
# server/package.json declares, into a directory the server's own
# module resolution walks. Verify the image this job just pushed, not
# a local build, so a build-cache or layer-ordering regression is
# caught before any tenant runs the image.
- name: Verify the pushed image resolves the declared Sentry version
env:
IMAGE: ghcr.io/${{ github.repository }}@${{ steps.build-cloud.outputs.digest }}
run: |
set -euo pipefail
expected="$(node -e "process.stdout.write(require('./server/package.json').peerDependencies['@sentry/node'])")"
test -n "$expected"
installed="$(docker run --rm --pull always \
-v "$PWD/scripts/assert-cloud-image-sentry.mjs:/app/server/.ci-sentry-probe.mjs:ro" \
--entrypoint node "$IMAGE" /app/server/.ci-sentry-probe.mjs)"
echo "Declared optional peer version: $expected"
echo "Installed in the pushed image: $installed"
if [ "$installed" != "$expected" ]; then
echo "ERROR: the pushed image resolves @sentry/node@$installed, expected @sentry/node@$expected" >&2
exit 1
fi
echo "The pushed image resolves the declared @sentry/node version."
# Verify the independently published cloud image without waiting for
# the self-hosted manifest job. The Sentry check already pulled it.
- name: Verify cloud PID 1 reaps orphaned processes
env:
IMAGE: ghcr.io/${{ github.repository }}@${{ steps.build-cloud.outputs.digest }}
run: docker run --rm -i "$IMAGE" sh -s < scripts/assert-orphan-reaping.sh
# Cloud's commit resolver and preview-artifact planner use the full SHA.
# Publish that address only after checking this build's exact digest.
# Retagging reuses the registry manifest and does not rebuild the image.
- name: Publish verified full-SHA cloud tag
env:
IMAGE: ghcr.io/${{ github.repository }}@${{ steps.build-cloud.outputs.digest }}
FULL_SHA_TAG: ghcr.io/${{ github.repository }}:sha-${{ github.sha }}-cloud
run: |
set -euo pipefail
revision="$(docker image inspect "$IMAGE" --format '{{ index .Config.Labels "org.opencontainers.image.revision" }}')"
platform="$(docker image inspect "$IMAGE" --format '{{ .Os }}/{{ .Architecture }}')"
test "$revision" = "$GITHUB_SHA"
test "$platform" = linux/amd64
docker buildx imagetools create --prefer-index=false --tag "$FULL_SHA_TAG" "$IMAGE"

View File

@ -349,8 +349,7 @@ jobs:
# until the cgroup pid limit is exhausted and every fork() in the
# container fails. Run against the pushed manifest rather than a local
# build: the legs push by digest, so nothing is loaded into this
# runner's daemon. The cloud variant is FROM production and inherits the
# same ENTRYPOINT, so checking this image covers both.
# runner's daemon. The independent cloud workflow checks its own image.
- name: Verify PID 1 reaps orphaned processes
env:
# Through the environment, not interpolated into the script body, so
@ -363,234 +362,15 @@ jobs:
echo "Verifying orphan reaping in $image"
docker run --rm -i --pull always "$image" sh -s < scripts/assert-orphan-reaping.sh
# The cloud variant carries built bundled plugins for managed deployments
# (see the `cloud` stage in the Dockerfile). It runs as its own job with no
# `needs:` on the stock publish above, so the two builds run in parallel and
# a failure or slow build in one never gates, delays, or skips the other.
# Both jobs share only the single top-level concurrency slot. Each job is a
# separate runner, so this one carries its own copy of the prep steps
# (checkout through schema labels) — the accepted cost of that isolation.
# Master cloud builds start independently in docker-cloud.yml. Tag builds
# and manual Docker dispatches call the same implementation, preserving the
# release tags and the canary promotion dependency below.
build-and-push-cloud:
runs-on: ubuntu-latest
timeout-minutes: 60
if: github.event_name != 'push' || github.ref != 'refs/heads/master'
uses: ./.github/workflows/docker-cloud.yml
permissions:
contents: read
packages: write
steps:
- name: Checkout
uses: actions/checkout@v7
with:
# Full history and tags so `git describe` below can compute the
# release version to stamp into the image.
fetch-depth: 0
# `.git` is dockerignored, so a running image cannot derive its own
# version and otherwise reports the source package.json placeholder in
# analytics and the debug panel. Compute it here from the pristine
# checkout (real CalVer drift from the nearest release tag) and pass it
# into the build. Empty when no release tag is reachable — the server
# then keeps its existing fallbacks.
- name: Compute build version
id: build-version
run: |
set -euo pipefail
case "${GITHUB_REF}" in
refs/tags/nightly/v*)
# Lane tags carry the exact published version; stamp it verbatim
# instead of describing drift from the nearest stable tag.
version="${GITHUB_REF#refs/tags/nightly/v}"
;;
refs/tags/beta/v*)
version="${GITHUB_REF#refs/tags/beta/v}"
;;
*)
version="$(git describe --tags --match 'v*' --long --dirty 2>/dev/null || true)"
;;
esac
echo "version=${version}" >> "$GITHUB_OUTPUT"
echo "Stamping build version: ${version:-<none>}"
# ISO week stamp for the Dockerfile's tool layer: the layer caches
# across commits and re-pulls the @latest CLI tools when the week rolls
# over, instead of on every build.
- name: Compute tool cache epoch
id: tools-epoch
run: echo "epoch=$(date -u +%G-W%V)" >> "$GITHUB_OUTPUT"
- name: Setup pnpm
uses: pnpm/action-setup@v6
with:
version: 9.15.4
run_install: false
# No dependency cache here: this workflow publishes release images, and
# restoring a shared Actions cache into the build inputs would let a
# poisoned cache entry reach the published artifact.
- name: Setup Node.js
uses: actions/setup-node@v7
with:
node-version: 24
- name: Refresh lockfile for Docker build context
run: |
set -euo pipefail
pnpm install --resolution-only --ignore-scripts --no-frozen-lockfile
changed="$(git status --porcelain)"
if [ -z "$changed" ]; then
echo "Lockfile already matches package metadata."
exit 0
fi
if printf '%s\n' "$changed" | grep -Fvq ' pnpm-lock.yaml'; then
echo "Unexpected files changed during lockfile refresh:"
echo "$changed"
exit 1
fi
echo "Using refreshed pnpm-lock.yaml in the Docker build context."
- name: Free runner disk
run: |
set -euo pipefail
echo "Disk before cleanup:"
df -h
pnpm store prune || true
sudo apt-get clean || true
sudo rm -rf \
/usr/share/dotnet \
/usr/share/swift \
/usr/local/lib/android \
/usr/local/share/boost \
/usr/local/share/powershell \
/opt/ghc \
/opt/hostedtoolcache/CodeQL \
/opt/hostedtoolcache/PyPy \
/opt/hostedtoolcache/Ruby || true
docker system prune -af || true
echo "Disk after cleanup:"
df -h
- name: Login to GitHub Container Registry
uses: docker/login-action@v4
with:
registry: ghcr.io
username: ${{ github.repository_owner }}
password: ${{ secrets.GITHUB_TOKEN }}
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v4
# Deployment tooling reads these labels from the registry to verify an
# image's schema expectations against a migrator before deploying it,
# without pulling the image. The server refuses to start when the
# database is missing bundled migrations, so orchestrators need a cheap
# way to check image/migrator compatibility up front.
- name: Compute schema migration labels
id: schema
run: |
set -euo pipefail
last=$(ls packages/db/src/migrations/*.sql | sed 's|.*/||' | LC_ALL=C sort | tail -1)
count=$(ls packages/db/src/migrations/*.sql | wc -l | tr -d ' ')
echo "last=${last}" >> "$GITHUB_OUTPUT"
echo "count=${count}" >> "$GITHUB_OUTPUT"
# Published under the same lane tag set as the self-hosted image, with a
# `-cloud` suffix (nightly-cloud, latest-cloud, <version>-cloud,
# sha-<short>-cloud). `:canary-cloud` follows the same retag-step
# ownership rule as `:canary` above.
- name: Docker meta (cloud)
id: meta-cloud
uses: docker/metadata-action@v6
with:
images: ghcr.io/${{ github.repository }}
flavor: |
suffix=-cloud,onlatest=true
tags: |
type=raw,value=nightly,enable=${{ startsWith(github.ref, 'refs/tags/nightly/v') }}
type=raw,value=beta,enable=${{ startsWith(github.ref, 'refs/tags/beta/v') }}
type=raw,value=latest,enable=${{ startsWith(github.ref, 'refs/tags/v') }}
type=semver,pattern={{version}},enable=${{ startsWith(github.ref, 'refs/tags/v') }}
type=semver,pattern={{major}}.{{minor}},enable=${{ startsWith(github.ref, 'refs/tags/v') }}
type=sha
labels: |
io.github.paperclipai.schema.last-migration=${{ steps.schema.outputs.last }}
io.github.paperclipai.schema.migration-count=${{ steps.schema.outputs.count }}
- name: Build and push (cloud)
id: build-cloud
uses: docker/build-push-action@v7
with:
context: .
target: cloud
# Space-separated sandbox-provider directory names to build into
# the variant; add here when managed deployments need another.
# CLOUD_BUNDLED_SERVER_DEPS names the optional peer packages the
# variant installs from server/package.json's declared version;
# add another name there when a managed tenant needs it.
build-args: |
CLOUD_BUNDLED_PLUGINS=daytona
CLOUD_BUNDLED_SERVER_DEPS=@sentry/node
PAPERCLIP_BUILD_VERSION=${{ steps.build-version.outputs.version }}
PAPERCLIP_BUILD_COMMIT=${{ github.sha }}
CLI_TOOLS_CACHE_EPOCH=${{ steps.tools-epoch.outputs.epoch }}
# amd64 only, unlike the self-hosted image above: the cloud variant
# is consumed exclusively by managed-deployment hosts, which run
# amd64. The QEMU-emulated arm64 half dominated this job's wall
# clock, and dropping it roughly halves time-to-deployable-image.
platforms: linux/amd64
push: true
# Registry-backed BuildKit cache, separate ref from the self-hosted
# job so the two parallel builds never clobber each other's cache
# manifest (see the rationale on the job above).
cache-from: type=registry,ref=ghcr.io/${{ github.repository }}:buildcache-cloud
cache-to: type=registry,ref=ghcr.io/${{ github.repository }}:buildcache-cloud,mode=max
tags: ${{ steps.meta-cloud.outputs.tags }}
labels: ${{ steps.meta-cloud.outputs.labels }}
# The cloud target installs @sentry/node at the version
# server/package.json declares, into a directory the server's own
# module resolution walks. Verify the image this job just pushed, not
# a local build, so a build-cache or layer-ordering regression is
# caught before any tenant runs the image.
- name: Verify the pushed image resolves the declared Sentry version
env:
IMAGE: ghcr.io/${{ github.repository }}@${{ steps.build-cloud.outputs.digest }}
run: |
set -euo pipefail
expected="$(node -e "process.stdout.write(require('./server/package.json').peerDependencies['@sentry/node'])")"
test -n "$expected"
installed="$(docker run --rm --pull always \
-v "$PWD/scripts/assert-cloud-image-sentry.mjs:/app/server/.ci-sentry-probe.mjs:ro" \
--entrypoint node "$IMAGE" /app/server/.ci-sentry-probe.mjs)"
echo "Declared optional peer version: $expected"
echo "Installed in the pushed image: $installed"
if [ "$installed" != "$expected" ]; then
echo "ERROR: the pushed image resolves @sentry/node@$installed, expected @sentry/node@$expected" >&2
exit 1
fi
echo "The pushed image resolves the declared @sentry/node version."
# Cloud's commit resolver and preview-artifact planner use the full SHA.
# Publish that address only after checking this build's exact digest.
# Retagging reuses the registry manifest and does not rebuild the image.
- name: Publish verified full-SHA cloud tag
env:
IMAGE: ghcr.io/${{ github.repository }}@${{ steps.build-cloud.outputs.digest }}
FULL_SHA_TAG: ghcr.io/${{ github.repository }}:sha-${{ github.sha }}-cloud
run: |
set -euo pipefail
revision="$(docker image inspect "$IMAGE" --format '{{ index .Config.Labels "org.opencontainers.image.revision" }}')"
platform="$(docker image inspect "$IMAGE" --format '{{ .Os }}/{{ .Architecture }}')"
test "$revision" = "$GITHUB_SHA"
test "$platform" = linux/amd64
docker buildx imagetools create --prefer-index=false --tag "$FULL_SHA_TAG" "$IMAGE"
# Moves the mutable `:canary` / `:canary-cloud` channel tags. Kept OUT
# of the build jobs and serialized in its own lane, and — the load-

View File

@ -27,7 +27,13 @@ docker build -t paperclip-local \
## Cloud image addresses
The Docker workflow publishes the managed deployment image for Linux AMD64.
After the pushed image passes its Sentry check, the workflow verifies its
`Docker cloud` starts on each master push independently of the multi-platform
self-hosted build. Different commits use separate concurrency groups and existing
GitHub-hosted runners, so an older production or cloud build does not hold the
new commit in a workflow queue. Available GitHub runner capacity still applies.
Release tags and manual `Docker` dispatches call the same cloud build workflow.
After the pushed image passes its Sentry and orphan-reaping checks, the workflow verifies its
commit label and platform and adds `ghcr.io/paperclipai/paperclip:sha-<full-commit-sha>-cloud`.
This address lets commit-based deployment tooling reuse the normal build.
Existing short-SHA and release tags remain available.

View File

@ -128,9 +128,25 @@ test("commits sharing a short prefix use separate full-SHA image addresses", asy
assert.deepEqual(urls.filter((url) => url.includes("/manifests/")), [sha, other].map((commit) => `https://ghcr.io/v2/paperclipai/paperclip/manifests/sha-${commit}-cloud`));
});
test("cloud builds start per commit and preserve tag promotion dependencies", () => {
const docker = readFileSync(new URL("../.github/workflows/docker.yml", import.meta.url), "utf8");
const cloud = readFileSync(new URL("../.github/workflows/docker-cloud.yml", import.meta.url), "utf8");
assert.match(cloud, /branches: \[master\]/);
assert.match(cloud, /workflow_call:/);
assert.match(cloud, /group: docker-cloud-\$\{\{ github.sha \}\}/);
assert.match(cloud, /cancel-in-progress: false/);
const caller = docker.split(" build-and-push-cloud:")[1].split(" promote_canary_channel:")[0];
assert.match(caller, /if: github.event_name != 'push' \|\| github.ref != 'refs\/heads\/master'/);
assert.match(caller, /uses: .\/.github\/workflows\/docker-cloud.yml/);
assert.match(docker.split(" promote_canary_channel:")[1], /needs: \[merge-and-push, build-and-push-cloud\]/);
const reaping = cloud.indexOf(" - name: Verify cloud PID 1 reaps orphaned processes");
assert.ok(reaping > cloud.indexOf(" - name: Verify the pushed image resolves the declared Sentry version"));
assert.ok(reaping < cloud.indexOf(" - name: Publish verified full-SHA cloud tag"));
});
test("normal cloud builds publish the checked digest only when source and platform match", () => {
const workflow = readFileSync(new URL("../.github/workflows/docker.yml", import.meta.url), "utf8");
const cloud = workflow.split(" build-and-push-cloud:")[1].split(" promote_canary_channel:")[0];
const workflow = readFileSync(new URL("../.github/workflows/docker-cloud.yml", import.meta.url), "utf8");
const cloud = workflow.split(" build-and-push-cloud:")[1];
const verify = cloud.indexOf(" - name: Verify the pushed image resolves the declared Sentry version");
const publish = cloud.indexOf(" - name: Publish verified full-SHA cloud tag");
assert.ok(verify >= 0 && publish > verify);

View File

@ -21,6 +21,7 @@ import { describe, expect, it } from "vitest";
const repoRoot = path.resolve(path.dirname(fileURLToPath(import.meta.url)), "..", "..", "..");
const dockerfile = readFileSync(path.join(repoRoot, "Dockerfile"), "utf8");
const workflow = readFileSync(path.join(repoRoot, ".github", "workflows", "docker.yml"), "utf8");
const cloudWorkflow = readFileSync(path.join(repoRoot, ".github", "workflows", "docker-cloud.yml"), "utf8");
/**
* Return the text of the Dockerfile stage that starts at the named target.
@ -49,7 +50,7 @@ describe("docker build-stamp wiring", () => {
});
it("passes PAPERCLIP_BUILD_COMMIT as a build-arg for both image targets", () => {
const argLines = [...workflow.matchAll(/^\s*PAPERCLIP_BUILD_COMMIT=.*$/gm)];
const argLines = [...`${workflow}\n${cloudWorkflow}`.matchAll(/^\s*PAPERCLIP_BUILD_COMMIT=.*$/gm)];
expect(
argLines.length,
"the docker workflow must pass PAPERCLIP_BUILD_COMMIT for the production and cloud builds",