Files
hyperframes/.github/workflows/lambda-arm64-image.yml
T
Miguel Ángel f6cbf87cdd feat(aws-lambda): arm64 video function image from the handler ZIP (#6064)
* feat(aws-lambda): arm64 video function image from the handler ZIP

The image is the release handler ZIP with its x86 ffmpeg and Chrome swapped for
the native ffmpeg tree and the sha256-pinned Sparticuz arm64 Chrome pack, so the
x86 and arm64 functions run the same handler code. The image stage only copies,
so any host builds it. CI builds it and runs ffmpeg, a real Chrome render and the
Lambda runtime on an ARM runner.

* fix(aws-lambda): launch Chrome in the image smoke the way Lambda does

@sparticuz/chromium inflates its AL2023 libraries only when AWS_EXECUTION_ENV
names a Node runtime; Lambda sets it, a bare docker run does not. The workflow
also runs on @sparticuz/chromium bumps and handler ZIP layout changes, which
re-stage the image.

* fix(aws-lambda): smoke the arm64 image through its own Lambda runtime

The smoke is now a handler invoked through the image's entrypoint, whose
bootstrap sets AWS_EXECUTION_ENV the way Lambda does, so @sparticuz/chromium
takes its Lambda branch with nothing set by hand. It returns the value it saw.

* fix(aws-lambda): agree the move of the ffmpeg build workflow into the image workflow

* fix(aws-lambda): build the Chrome pack URL from the pin and hand the checked pack to tar in memory
2026-10-23 02:41:18 +00:00

107 lines
4.1 KiB
YAML

name: Lambda arm64 image
permissions:
contents: read
# Builds the arm64 Lambda video function image: the native ffmpeg on two separate ARM runners (each runs
# the codec check, the tarballs must match), then the image from the x86 handler ZIP, smoke-run on ARM.
on:
pull_request:
paths:
- "packages/aws-lambda/image/**"
- "packages/aws-lambda/scripts/stage-arm64-image.ts"
# A @sparticuz/chromium bump or a ZIP layout change must re-stage the image.
- "packages/aws-lambda/package.json"
- "packages/aws-lambda/scripts/build-zip.ts"
- ".github/workflows/lambda-arm64-image.yml"
workflow_dispatch:
concurrency:
group: lambda-arm64-image-${{ github.ref }}
cancel-in-progress: true
jobs:
ffmpeg:
name: ffmpeg ${{ matrix.copy }}
runs-on: ubuntu-24.04-arm
timeout-minutes: 90
strategy:
matrix:
copy: [a, b]
steps:
- uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4
- name: Build and check
run: >-
docker buildx build --progress=plain --platform linux/arm64 --target artifact
--output type=local,dest=out packages/aws-lambda/image
- run: cat out/hf-media-check.txt && sha256sum out/*.tar.gz
- uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4
with:
name: hf-media-${{ matrix.copy }}
path: out/
retention-days: 7
reproducible:
name: ffmpeg builds match
needs: ffmpeg
runs-on: ubuntu-latest
timeout-minutes: 5
steps:
- uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4
- run: cmp hf-media-a/hf-media-linux-aarch64.tar.gz hf-media-b/hf-media-linux-aarch64.tar.gz
handler-zip:
name: Handler ZIP
runs-on: ubuntu-latest
timeout-minutes: 30
steps:
- uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4
- uses: oven-sh/setup-bun@0c5077e51419868618aeaa5fe8019c62421857d6 # v2
- run: bun install --frozen-lockfile
- run: bun run build
- run: bun run --cwd packages/aws-lambda build:zip
- uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4
with:
name: handler-zip
path: packages/aws-lambda/dist/handler.zip
retention-days: 7
image:
name: Image smoke
needs: [ffmpeg, handler-zip]
runs-on: ubuntu-24.04-arm
timeout-minutes: 20
steps:
- uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4
- uses: oven-sh/setup-bun@0c5077e51419868618aeaa5fe8019c62421857d6 # v2
- uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4
with:
path: artifacts
- name: Stage /var/task
run: >-
bun packages/aws-lambda/scripts/stage-arm64-image.ts
--zip=artifacts/handler-zip/handler.zip
--ffmpeg=artifacts/hf-media-a/hf-media-linux-aarch64.tar.gz --out=staged
- name: Build image
run: >-
docker buildx build --progress=plain --platform linux/arm64 --target video
--build-context task=staged/task --load -t hf-video:ci packages/aws-lambda/image
- name: ffmpeg and Chrome run under the image's own Lambda runtime
run: |
docker run -d --name hf-smoke -p 9001:8080 \
-v "$PWD/packages/aws-lambda/image/smoke-video.mjs:/var/task/smoke-video.mjs:ro" \
hf-video:ci smoke-video.handler
reply=$(curl -sS --retry 30 --retry-all-errors --retry-delay 1 \
-d '{}' http://localhost:9001/2015-03-31/functions/function/invocations)
docker logs hf-smoke
echo "$reply"
jq -e '.arch == "arm64" and .screenshotBytes > 0 and (.ffmpeg | startswith("ffmpeg version 7.0.2"))' <<<"$reply"
- name: The Lambda runtime loads the handler
run: |
docker run -d --name hf-video -p 9000:8080 hf-video:ci
reply=$(curl -sS --retry 30 --retry-all-errors --retry-delay 1 \
-d '{"Action":"not-an-action"}' http://localhost:9000/2015-03-31/functions/function/invocations)
docker logs hf-video
echo "$reply"
grep -q 'no recognised Action' <<<"$reply"