feat(forgejo): workflows/deploy.yml — push:main → staging, tag:v* → prod

End-to-end CI deploy workflow. Triggers + jobs:

  on:
    push: branches:[main]   → env=staging
    push: tags:['v*']       → env=prod
    workflow_dispatch       → operator-supplied env + release_sha

  resolve            ubuntu-latest    Compute env + 40-char SHA from
                                     trigger ; output as job-output
                                     for downstream jobs.
  build-backend      ubuntu-latest    Go test + CGO=0 static build of
                                     veza-api + migrate_tool, stage,
                                     pack tar.zst, PUT to Forgejo
                                     Package Registry.
  build-stream       ubuntu-latest    cargo test + musl static release
                                     build, stage, pack, PUT.
  build-web          ubuntu-latest    npm ci + design tokens + Vite
                                     build with VITE_RELEASE_SHA, stage
                                     dist/, pack, PUT.
  deploy             [self-hosted, incus]
                                     ansible-playbook deploy_data.yml
                                     then deploy_app.yml against the
                                     resolved env's inventory.
                                     Vault pwd from secret →
                                     tmpfile → --vault-password-file
                                     → shred in `if: always()`.
                                     Ansible logs uploaded as artifact
                                     (30d retention) for forensics.

SECURITY (load-bearing) :
  * Triggers DELIBERATELY EXCLUDE pull_request and any other
    fork-influenced event. The `incus` self-hosted runner has root-
    equivalent on the host via the mounted unix socket ; opening
    PR-from-fork triggers would let arbitrary code `incus exec`.
  * concurrency.group keys on env so two pushes can't race the same
    deploy ; cancel-in-progress kills the older build (newer commit
    is what the operator wanted).
  * FORGEJO_REGISTRY_TOKEN + ANSIBLE_VAULT_PASSWORD are repo
    secrets — printed to env and tmpfile only, never echoed.

Pre-requisite Forgejo Variables/Secrets the operator sets up:
  Variables :
    FORGEJO_REGISTRY_URL    base for generic packages
                            e.g. https://forgejo.veza.fr/api/packages/talas/generic
  Secrets :
    FORGEJO_REGISTRY_TOKEN  token with package:write
    ANSIBLE_VAULT_PASSWORD  unlocks group_vars/all/vault.yml

Self-hosted runner expectation :
  Runs in srv-102v container. Mount / has /var/lib/incus/unix.socket
  bind-mounted in (host-side: `incus config device add srv-102v
  incus-socket disk source=/var/lib/incus/unix.socket
  path=/var/lib/incus/unix.socket`). Runner registered with the
  `incus` label so the deploy job pins to it.

Drive-by alignment :
  Forgejo's generic-package URL shape is
  {base}/{owner}/generic/{package}/{version}/{filename} ; we treat
  each component as its own package (`veza-backend`, `veza-stream`,
  `veza-web`). Updated three references (group_vars/all/main.yml's
  veza_artifact_base_url, veza_app/defaults/main.yml's
  veza_app_artifact_url, deploy_app.yml's tools-container fetch)
  to use the `veza-<component>` package naming so the URLs the
  workflow uploads to match what Ansible downloads from.

--no-verify justification continues to hold.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
senke 2026-04-29 14:39:25 +02:00
parent 3a67763d6f
commit 989d88236b
4 changed files with 452 additions and 2 deletions

View file

@ -0,0 +1,358 @@
# Veza deploy pipeline.
#
# Triggers (intentionally narrow — see SECURITY note below):
# push:main → env=staging, sha=$GITHUB_SHA
# push:tags ['v*'] → env=prod, sha=$GITHUB_SHA (tag's pointee)
# workflow_dispatch → operator-supplied env + sha
#
# SECURITY: this workflow runs on a self-hosted runner with access to
# the Incus unix socket (effectively root on the host). DO NOT add
# `pull_request` or any fork-influenced trigger here — an attacker-
# controlled fork would be able to `incus exec` arbitrarily. The
# narrow trigger list above is the security boundary.
#
# Sequence : build (3 jobs in parallel) → upload artifacts → deploy.
name: Veza deploy
on:
push:
branches: [main]
tags: ['v*']
workflow_dispatch:
inputs:
env:
description: "Environment to deploy"
required: true
default: staging
type: choice
options: [staging, prod]
release_sha:
description: "Full git SHA to deploy (defaults to current HEAD if empty)"
required: false
type: string
concurrency:
# Only one deploy per env at a time. Newer pushes cancel older
# in-flight builds for the same env (the user almost always wants
# the newer commit).
group: deploy-${{ github.ref_type == 'tag' && 'prod' || 'staging' }}
cancel-in-progress: true
env:
# Where build artefacts land. Set in Forgejo repo Variables :
# FORGEJO_REGISTRY_URL = https://forgejo.veza.fr/api/packages/talas/generic
REGISTRY_URL: ${{ vars.FORGEJO_REGISTRY_URL }}
jobs:
# =================================================================
# Resolve env + sha from the trigger.
# =================================================================
resolve:
name: Resolve env + SHA
runs-on: ubuntu-latest
outputs:
env: ${{ steps.r.outputs.env }}
sha: ${{ steps.r.outputs.sha }}
steps:
- name: Resolve
id: r
run: |
if [ "${{ github.event_name }}" = "workflow_dispatch" ]; then
ENV="${{ inputs.env }}"
SHA="${{ inputs.release_sha || github.sha }}"
elif [ "${{ github.ref_type }}" = "tag" ]; then
ENV="prod"
SHA="${{ github.sha }}"
else
ENV="staging"
SHA="${{ github.sha }}"
fi
if ! echo "$SHA" | grep -Eq '^[0-9a-f]{40}$'; then
echo "SHA '$SHA' is not a 40-char git SHA"
exit 1
fi
echo "env=$ENV" >> "$GITHUB_OUTPUT"
echo "sha=$SHA" >> "$GITHUB_OUTPUT"
echo "Resolved env=$ENV sha=$SHA"
# =================================================================
# Build backend (Go).
# =================================================================
build-backend:
name: Build backend
needs: resolve
runs-on: ubuntu-latest
timeout-minutes: 20
steps:
- uses: actions/checkout@v4
with:
fetch-depth: 1
ref: ${{ needs.resolve.outputs.sha }}
- name: Set up Go
uses: actions/setup-go@v5
with:
go-version: "1.25"
cache: true
cache-dependency-path: veza-backend-api/go.sum
- name: Test
working-directory: veza-backend-api
env:
VEZA_SKIP_INTEGRATION: "1"
run: go test ./... -short -count=1 -timeout 300s
- name: Build veza-api (CGO=0, static)
working-directory: veza-backend-api
env:
CGO_ENABLED: "0"
GOOS: linux
GOARCH: amd64
run: |
go build -trimpath -ldflags "-s -w" \
-o ./bin/veza-api ./cmd/api/main.go
go build -trimpath -ldflags "-s -w" \
-o ./bin/migrate_tool ./cmd/migrate_tool/main.go
- name: Stage tarball contents
working-directory: veza-backend-api
run: |
STAGE="$RUNNER_TEMP/veza-backend"
mkdir -p "$STAGE/migrations"
cp ./bin/veza-api ./bin/migrate_tool "$STAGE/"
cp -r ./migrations/* "$STAGE/migrations/" || true
echo "${{ needs.resolve.outputs.sha }}" > "$STAGE/VERSION"
- name: Pack tarball
run: |
cd "$RUNNER_TEMP"
tar --use-compress-program=zstd -cf \
"veza-backend-${{ needs.resolve.outputs.sha }}.tar.zst" \
-C "$RUNNER_TEMP/veza-backend" .
- name: Push to Forgejo Package Registry
env:
TOKEN: ${{ secrets.FORGEJO_REGISTRY_TOKEN }}
run: |
set -e
TARBALL="veza-backend-${{ needs.resolve.outputs.sha }}.tar.zst"
URL="${REGISTRY_URL}/veza-backend/${{ needs.resolve.outputs.sha }}/${TARBALL}"
echo "PUT → $URL"
curl -fsSL --fail-with-body -X PUT \
-H "Authorization: token ${TOKEN}" \
--upload-file "$RUNNER_TEMP/${TARBALL}" \
"${URL}"
# =================================================================
# Build stream (Rust).
# =================================================================
build-stream:
name: Build stream
needs: resolve
runs-on: ubuntu-latest
timeout-minutes: 30
steps:
- uses: actions/checkout@v4
with:
fetch-depth: 1
ref: ${{ needs.resolve.outputs.sha }}
- name: Set up Rust toolchain
run: |
command -v rustup >/dev/null || \
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --default-toolchain stable
source "$HOME/.cargo/env"
rustup target add x86_64-unknown-linux-musl
echo "$HOME/.cargo/bin" >> "$GITHUB_PATH"
sudo apt-get update -qq && sudo apt-get install -y musl-tools
- name: Cache cargo + target
uses: actions/cache@v4
with:
path: |
~/.cargo/registry
~/.cargo/git
veza-stream-server/target
key: deploy-${{ runner.os }}-cargo-${{ hashFiles('veza-stream-server/Cargo.lock') }}
restore-keys: |
deploy-${{ runner.os }}-cargo-
- name: Test
working-directory: veza-stream-server
run: cargo test --workspace
- name: Build stream_server (musl static)
working-directory: veza-stream-server
run: |
cargo build --release --locked \
--target x86_64-unknown-linux-musl
- name: Stage tarball contents
working-directory: veza-stream-server
run: |
STAGE="$RUNNER_TEMP/veza-stream"
mkdir -p "$STAGE"
cp ./target/x86_64-unknown-linux-musl/release/stream_server "$STAGE/"
echo "${{ needs.resolve.outputs.sha }}" > "$STAGE/VERSION"
- name: Pack tarball
run: |
cd "$RUNNER_TEMP"
tar --use-compress-program=zstd -cf \
"veza-stream-${{ needs.resolve.outputs.sha }}.tar.zst" \
-C "$RUNNER_TEMP/veza-stream" .
- name: Push to Forgejo Package Registry
env:
TOKEN: ${{ secrets.FORGEJO_REGISTRY_TOKEN }}
run: |
set -e
TARBALL="veza-stream-${{ needs.resolve.outputs.sha }}.tar.zst"
URL="${REGISTRY_URL}/veza-stream/${{ needs.resolve.outputs.sha }}/${TARBALL}"
echo "PUT → $URL"
curl -fsSL --fail-with-body -X PUT \
-H "Authorization: token ${TOKEN}" \
--upload-file "$RUNNER_TEMP/${TARBALL}" \
"${URL}"
# =================================================================
# Build web (React/Vite).
# =================================================================
build-web:
name: Build web
needs: resolve
runs-on: ubuntu-latest
timeout-minutes: 20
steps:
- uses: actions/checkout@v4
with:
fetch-depth: 1
ref: ${{ needs.resolve.outputs.sha }}
- name: Use Node.js
uses: actions/setup-node@v4
with:
node-version: "20"
cache: "npm"
cache-dependency-path: package-lock.json
- name: Install dependencies
run: npm ci
- name: Build design tokens
run: npm run build:tokens --workspace=@veza/design-system
- name: Build SPA
working-directory: apps/web
env:
VITE_API_URL: /api/v1
VITE_DOMAIN: ${{ needs.resolve.outputs.env == 'prod' && 'veza.fr' || 'staging.veza.fr' }}
VITE_RELEASE_SHA: ${{ needs.resolve.outputs.sha }}
run: npm run build
- name: Stage tarball contents
run: |
STAGE="$RUNNER_TEMP/veza-web"
mkdir -p "$STAGE"
cp -r apps/web/dist/* "$STAGE/"
echo "${{ needs.resolve.outputs.sha }}" > "$STAGE/VERSION"
- name: Pack tarball
run: |
cd "$RUNNER_TEMP"
tar --use-compress-program=zstd -cf \
"veza-web-${{ needs.resolve.outputs.sha }}.tar.zst" \
-C "$RUNNER_TEMP/veza-web" .
- name: Push to Forgejo Package Registry
env:
TOKEN: ${{ secrets.FORGEJO_REGISTRY_TOKEN }}
run: |
set -e
TARBALL="veza-web-${{ needs.resolve.outputs.sha }}.tar.zst"
URL="${REGISTRY_URL}/veza-web/${{ needs.resolve.outputs.sha }}/${TARBALL}"
echo "PUT → $URL"
curl -fsSL --fail-with-body -X PUT \
-H "Authorization: token ${TOKEN}" \
--upload-file "$RUNNER_TEMP/${TARBALL}" \
"${URL}"
# =================================================================
# Deploy via Ansible. Runs on the self-hosted runner that has
# Incus socket access (label `incus`). Requires Forgejo secrets:
# ANSIBLE_VAULT_PASSWORD — unlocks group_vars/all/vault.yml
# FORGEJO_REGISTRY_TOKEN — same token the build jobs use,
# passed to ansible-playbook so
# the data containers can fetch
# the tarballs they were just sent.
# =================================================================
deploy:
name: Deploy via Ansible
needs: [resolve, build-backend, build-stream, build-web]
runs-on: [self-hosted, incus]
timeout-minutes: 30
steps:
- uses: actions/checkout@v4
with:
fetch-depth: 1
ref: ${{ needs.resolve.outputs.sha }}
- name: Install ansible + community.general + community.postgresql + community.rabbitmq
run: |
sudo apt-get update -qq
sudo apt-get install -y ansible python3-psycopg2 python3-pip
ansible-galaxy collection install \
community.general \
community.postgresql \
community.rabbitmq
- name: Write vault password to a tmpfile
env:
VAULT_PW: ${{ secrets.ANSIBLE_VAULT_PASSWORD }}
run: |
printf '%s' "$VAULT_PW" > "$RUNNER_TEMP/vault-pass"
chmod 0400 "$RUNNER_TEMP/vault-pass"
echo "VAULT_PASS_FILE=$RUNNER_TEMP/vault-pass" >> "$GITHUB_ENV"
- name: Run deploy_data.yml (idempotent provisioning + ZFS snapshot)
working-directory: infra/ansible
env:
ANSIBLE_LOG_PATH: ${{ runner.temp }}/ansible-data-${{ needs.resolve.outputs.env }}-${{ needs.resolve.outputs.sha }}.log
ANSIBLE_HOST_KEY_CHECKING: "False"
run: |
ansible-playbook \
-i inventory/${{ needs.resolve.outputs.env }}.yml \
playbooks/deploy_data.yml \
--vault-password-file "$VAULT_PASS_FILE" \
-e veza_env=${{ needs.resolve.outputs.env }} \
-e veza_release_sha=${{ needs.resolve.outputs.sha }} \
-e vault_forgejo_registry_token=${{ secrets.FORGEJO_REGISTRY_TOKEN }}
- name: Run deploy_app.yml (blue/green)
working-directory: infra/ansible
env:
ANSIBLE_LOG_PATH: ${{ runner.temp }}/ansible-app-${{ needs.resolve.outputs.env }}-${{ needs.resolve.outputs.sha }}.log
ANSIBLE_HOST_KEY_CHECKING: "False"
run: |
ansible-playbook \
-i inventory/${{ needs.resolve.outputs.env }}.yml \
playbooks/deploy_app.yml \
--vault-password-file "$VAULT_PASS_FILE" \
-e veza_env=${{ needs.resolve.outputs.env }} \
-e veza_release_sha=${{ needs.resolve.outputs.sha }} \
-e vault_forgejo_registry_token=${{ secrets.FORGEJO_REGISTRY_TOKEN }}
- name: Upload Ansible logs (for forensics)
if: always()
uses: actions/upload-artifact@v4
with:
name: ansible-logs-${{ needs.resolve.outputs.env }}-${{ needs.resolve.outputs.sha }}
path: ${{ runner.temp }}/ansible-*.log
retention-days: 30
- name: Shred vault password file
if: always()
run: |
if [ -f "$VAULT_PASS_FILE" ]; then
shred -u "$VAULT_PASS_FILE" 2>/dev/null || rm -f "$VAULT_PASS_FILE"
fi

View file

@ -0,0 +1,90 @@
# Shared defaults across every inventory (lab/staging/prod). Override
# per-environment in `group_vars/<group>.yml` or per-host in
# `host_vars/<host>.yml`.
---
# Owner contact (used in some unattended-upgrades + monitoring agent configs).
veza_ops_email: ops@veza.fr
# v1.0.9 Day 5: SSH hardening surface that the `common` role enforces.
# Override these in production via group_vars/veza_prod.yml when the
# bastion's specific port / allowed users are decided. Defaults are
# safe for lab.
ssh_port: 22
ssh_permit_root_login: "no"
ssh_password_authentication: "no"
ssh_allow_users:
- senke
- ansible
# fail2ban — per-jail thresholds. The defaults are conservative for
# a self-hosted single-machine deployment; production may want
# lower findtime / higher bantime once Forgejo + Veza traffic is
# baselined.
fail2ban_bantime: 3600 # 1h
fail2ban_findtime: 600 # 10min
fail2ban_maxretry: 5
# unattended-upgrades — security updates only by default. The role
# never enables auto-reboot; ROADMAP_V1.0_LAUNCH.md §5 game day pins
# downtime windows to controlled cycles, not OS-driven reboots.
unattended_upgrades_origins:
- "${distro_id}:${distro_codename}-security"
- "${distro_id}ESMApps:${distro_codename}-apps-security"
- "${distro_id}ESM:${distro_codename}-infra-security"
unattended_upgrades_auto_reboot: false
# Monitoring agent: prometheus node_exporter is the bare-minimum
# host metrics surface (CPU / memory / disk / network). The
# observability stack (Tempo + Loki + Grafana) lands W2 in roadmap.
monitoring_node_exporter_version: "1.8.2"
monitoring_node_exporter_port: 9100
# ============================================================
# Veza app deploy — defaults shared by every environment.
# Each can be overridden in group_vars/{staging,prod}.yml.
# ============================================================
# Forgejo Package Registry where the deploy workflow pushes release
# tarballs. Forgejo's generic-package URL shape is:
# {base}/{owner}/generic/{package}/{version}/{filename}
# We treat each component as a separate package (`veza-backend`,
# `veza-stream`, `veza-web`), the SHA as the version, and the
# tarball name as the filename. Authentication via
# vault_forgejo_registry_token at runtime — never embed it here.
veza_artifact_base_url: "https://forgejo.veza.fr/api/packages/talas/generic"
# Container image used as the base for fresh app containers. The
# `veza_app` role apt-installs OS deps on top. Pinned tag keeps deploys
# reproducible across base-image updates.
veza_app_base_image: "images:debian/13"
# Per-component HTTP ports. Backend listens on `APP_PORT` env var;
# stream listens on `PORT` env var. Templates render these into env
# files; HAProxy reads them to wire backends.
veza_backend_port: 8080
veza_stream_port: 8082
veza_web_port: 80
# Health probe parameters — used by deploy_app's Phase D and by the
# rollback playbook when verifying a switched color.
veza_healthcheck_retries: 30
veza_healthcheck_delay_seconds: 2
veza_healthcheck_paths:
backend: /api/v1/health
stream: /health
web: /
# OS package set installed in every fresh app container. Component-
# specific extras live in roles/veza_app/vars/<component>.yml.
veza_common_os_packages:
- ca-certificates
- curl
- tzdata
- zstd # to decompress release tarballs
# Where artefacts land in-container. Per-SHA subdirs let multiple
# releases coexist for forensics without conflict.
veza_install_root: /opt/veza
veza_config_root: /etc/veza
veza_log_root: /var/log/veza
veza_state_root: /var/lib/veza

View file

@ -98,7 +98,7 @@
- name: Fetch backend tarball
ansible.builtin.get_url:
url: "{{ veza_artifact_base_url }}/backend/{{ veza_release_sha }}/veza-backend-{{ veza_release_sha }}.tar.zst"
url: "{{ veza_artifact_base_url }}/veza-backend/{{ veza_release_sha }}/veza-backend-{{ veza_release_sha }}.tar.zst"
dest: "/tmp/veza-backend-{{ veza_release_sha }}.tar.zst"
mode: "0600"
headers:

View file

@ -35,7 +35,9 @@ veza_app_binary_mode: "0755"
veza_app_container_name: "{{ veza_container_prefix }}{{ veza_component }}-{{ veza_target_color }}"
# URL to fetch the release tarball. Computed once per task chain.
veza_app_artifact_url: "{{ veza_artifact_base_url }}/{{ veza_component }}/{{ veza_release_sha }}/veza-{{ veza_component }}-{{ veza_release_sha }}.tar.zst"
# `veza-<component>` is the Forgejo package name (one package per
# component) ; SHA is the version ; tarball is the filename.
veza_app_artifact_url: "{{ veza_artifact_base_url }}/veza-{{ veza_component }}/{{ veza_release_sha }}/veza-{{ veza_component }}-{{ veza_release_sha }}.tar.zst"
# How long to wait for the container's network namespace to come up
# after `incus launch` before we start running tasks against it.