open-standards/.github/workflows/wai-fleet.yml
Transaction Science 6517401f01 CI: the gate that decides whether a workflow runs was queued behind the work it gates
Every path-gated workflow begins with one small job -- `does this change touch
it`, or `which workspaces this change touches` for the standards build -- that
all of its real jobs declare in `needs`. All 55 of those gate jobs ran on
`linux-arm64`.

That is the one label the fleet cannot spare. Sampled while 267 runs were
queued, every waiting run had the same shape:

    [waiting]  linux-arm64    does this change touch it
    [blocked]  linux          member · dataflow device, simulated
    [blocked]  macos          member · dataflow device absent

237 runs, each holding a gate in the longest queue, and every real job blocked
behind it. Not only `linux` jobs: `macos`, `linux-hw`, `linux-arm64-hw` and
`macos-m3-ultra` jobs were blocked too. The measured effect was `linux-arm64`
263 deep against `linux` **zero**, with jetson-hub's 20 cores taking no job for
a full hour -- starved not for want of work but because the work could not be
unblocked. A gate waiting on the resource it is rationing is head-of-line
blocking, and it makes every other runner in the fleet idle on its behalf.

The gates move to `linux`. Nothing else changes: the diff is 55 `runs-on` lines
and no others, and every touched file still parses.

Why `linux` and not elsewhere. The gate is not free -- it compares
`BASE...HEAD`, which needs the merge base, so `fetch-depth: 0` is a full clone
rather than a ten-second script. `macos` is served by hosts on a Mac Studio
sitting at load 185; `linux` is a separate 20-core box that was measurably
empty. This moves a serialisation point to the least contended label; it does
not remove it, and the gate is still a single point of failure for its
workflow. The structural fix is an early-exit first step inside each job, which
costs one checkout per leg and is a larger change.

Follow-up, deliberately not bundled: jetson-hub runs at capacity 2 with 15 GB of
RAM, so 55 clones still serialise two at a time. Raising capacity helps gates
and hurts the heavy emulated-architecture jobs that share the label, so it wants
measuring on its own.

Co-Authored-By: Claude Opus 5 <[email protected]>
2026-11-03 15:19:13 -05:00

130 lines
5.6 KiB
YAML

name: wai-fleet service
# A WAI fleet as a service (wai/wai-fleet, extensions/fleet.md §9): its tests (the corpus's report
# signed byte for byte, the directory recovered after a restart, a line a crash cut short cut off,
# one process per directory, every summary of the corpus read and written again by the reference
# relay's own types, and records the relay's own code signs collected from a file over HTTP),
# clippy with warnings as errors, and its smoke checks over HTTP against an instance started here
# with the corpus's organisation key, the report's bundle checked by fleet_verify.py. On Linux;
# on the macOS runner (a host the service would be deployed to) the release build and the smoke
# checks, without the tests, which build the reference relay.
concurrency:
group: ${{ github.workflow }}-${{ github.ref }}
cancel-in-progress: true
on:
push:
branches: [main]
paths:
- 'wai/wai-rs/**'
- 'wai/wai-fleet/**'
- 'wai/live-conformance/**'
- 'wai/claim-conformance/**'
- 'wai/extensions/fleet.md'
- 'joule-code/joule-code-rs/crates/joule-moq-relay/**'
- 'joule-code/joule-code-rs/crates/joule-moq/**'
- 'joule-code/joule-code-rs/crates/jwp/**'
- 'joule-code/joule-code-rs/vendor/wtransport/**'
- 'rust-toolchain.toml'
- '.github/workflows/wai-fleet.yml'
pull_request:
paths:
- 'wai/wai-rs/**'
- 'wai/wai-fleet/**'
- 'wai/live-conformance/**'
- 'wai/claim-conformance/**'
- 'wai/extensions/fleet.md'
- 'joule-code/joule-code-rs/crates/joule-moq-relay/**'
- 'joule-code/joule-code-rs/crates/joule-moq/**'
- 'joule-code/joule-code-rs/crates/jwp/**'
- 'joule-code/joule-code-rs/vendor/wtransport/**'
- 'rust-toolchain.toml'
- '.github/workflows/wai-fleet.yml'
jobs:
relevant:
name: does this change touch it
runs-on: linux
outputs:
run: ${{ steps.gate.outputs.run }}
steps:
- uses: actions/checkout@v5
with:
fetch-depth: 0
- id: gate
run: |
set -euo pipefail
r=$(python3 scripts/pr-touches-paths.py .github/workflows/wai-fleet.yml \
"${{ github.event.pull_request.base.sha }}" \
"${{ github.event.pull_request.head.sha }}")
echo "this change touches wai-fleet: $r"
echo "run=$r" >> "$GITHUB_OUTPUT"
service-linux:
name: service · linux
needs: relevant
if: needs.relevant.outputs.run == 'true'
runs-on: linux
timeout-minutes: 60
steps:
- uses: actions/checkout@v5
- name: Install Rust 1.98.0
shell: bash
run: |
export PATH="$HOME/.cargo/bin:$PATH"
command -v rustup >/dev/null 2>&1 || \
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --profile minimal --default-toolchain 1.98.0
rustup toolchain install 1.98.0 --profile minimal --component clippy
rustup default 1.98.0
echo "$HOME/.cargo/bin" >> "$GITHUB_PATH"
- name: Tests and clippy
shell: bash
working-directory: wai/wai-fleet
run: |
set -euo pipefail
cargo test --locked --release 2>&1 | tee "$RUNNER_TEMP/test.log"
grep -qE '^test result: ok\. [1-9][0-9]* passed; 0 failed' "$RUNNER_TEMP/test.log" || { echo "::error::no test ran"; exit 1; }
cargo clippy --locked --release --all-targets -- -D warnings
- name: The smoke checks over HTTP
shell: bash
working-directory: wai/wai-fleet
run: |
set -euo pipefail
cargo build --locked --release
dir="$RUNNER_TEMP/fleet"; key="$RUNNER_TEMP/organisation.key"; rm -rf "$dir" "$key"
# The corpus's organisation key, from its fixed seed (a test key, never a deployment's).
python3 -I -c 'import hashlib,sys; sys.stdout.buffer.write(hashlib.sha256(b"wai:fleet/corpus/key\x00organisation").digest())' > "$key"
WAI_FLEET_KEY="$key" WAI_FLEET_DIR="$dir" WAI_FLEET_ADDR=127.0.0.1:18252 \
"${CARGO_TARGET_DIR:-target}/release/wai-fleet" serve 2> "$RUNNER_TEMP/service.log" &
pid=$!
trap 'kill $pid 2>/dev/null || true; cat "$RUNNER_TEMP/service.log"' EXIT
for _ in $(seq 1 50); do curl -s -o /dev/null http://127.0.0.1:18252/v1/fleet/health && break; sleep 0.2; done
bash smoke.sh http://127.0.0.1:18252
service-macos:
name: service · macos
needs: relevant
if: needs.relevant.outputs.run == 'true'
runs-on: macos
timeout-minutes: 60
steps:
- uses: actions/checkout@v5
# launchd's PATH omits cargo on the macOS host runner.
- name: Put cargo on PATH
shell: bash
run: echo "$HOME/.cargo/bin" >> "$GITHUB_PATH"
- name: The smoke checks over HTTP
shell: bash
working-directory: wai/wai-fleet
run: |
set -euo pipefail
nice -n 10 cargo build --locked --release
dir="$RUNNER_TEMP/fleet"; key="$RUNNER_TEMP/organisation.key"; rm -rf "$dir" "$key"
python3 -I -c 'import hashlib,sys; sys.stdout.buffer.write(hashlib.sha256(b"wai:fleet/corpus/key\x00organisation").digest())' > "$key"
WAI_FLEET_KEY="$key" WAI_FLEET_DIR="$dir" WAI_FLEET_ADDR=127.0.0.1:18252 \
"${CARGO_TARGET_DIR:-target}/release/wai-fleet" serve 2> "$RUNNER_TEMP/service.log" &
pid=$!
trap 'kill $pid 2>/dev/null || true; cat "$RUNNER_TEMP/service.log"' EXIT
for _ in $(seq 1 50); do curl -s -o /dev/null http://127.0.0.1:18252/v1/fleet/health && break; sleep 0.2; done
bash smoke.sh http://127.0.0.1:18252