Fam Zheng
857c0d5481
deploy articulate / build-and-deploy (push) Successful in 1m29s
deploy cube / build-and-deploy (push) Successful in 1m49s
deploy karaoke / build-and-deploy (push) Successful in 1m18s
deploy llm-proxy / build-and-deploy (push) Successful in 2m41s
deploy music / build-and-deploy (push) Successful in 3m6s
deploy notes / build-and-deploy (push) Successful in 2m40s
deploy simpleasm / build-and-deploy (push) Successful in 2m5s
deploy werewolf / build-and-deploy (push) Successful in 1m41s
新 service，ns `llm-proxy`，域 `llm.famzheng.me`。 - POST /v1/chat/completions — OpenAI 兼容透传到 mochi 同款 backend gateway (gemma-4-31b-it)；一期强制 stream=false，SSE 留二期 - 鉴权: `Authorization: token <PROXY_AUTH_TOKEN>` 或同款 Bearer； 常时间比较防 timing；空 expected 一律拒 - GET /chat — 自带极简 HTML chat UI（token 走 localStorage， 附 curl example details）；/ 跳转到 /chat - Secrets `llm-proxy/proxy-credentials` 已 kubectl 手工创建： BACKEND_TOKEN (上游) + PROXY_AUTH_TOKEN (对外) - 13 个 cargo test 覆盖 auth 多个 scheme / 边界 + body 改写 (stream=false 强制注入)
53 lines
1.6 KiB
YAML
53 lines
1.6 KiB
YAML
name: deploy llm-proxy
|
||
# llm.famzheng.me — gemma 反向代理。host shell runner（fam 用户）。
|
||
|
||
on:
|
||
push:
|
||
branches: [master]
|
||
paths:
|
||
- 'apps/llm-proxy/**'
|
||
- 'crates/cube-core/**'
|
||
- 'Cargo.toml'
|
||
- 'Cargo.lock'
|
||
- '.gitea/workflows/deploy-llm-proxy.yml'
|
||
workflow_dispatch:
|
||
|
||
jobs:
|
||
build-and-deploy:
|
||
runs-on: ubuntu-latest
|
||
env:
|
||
APP: llm-proxy
|
||
IMAGE: registry.famzheng.me/mochi/llm-proxy
|
||
steps:
|
||
- uses: actions/checkout@v4
|
||
|
||
- name: Resolve image tag
|
||
id: tag
|
||
run: echo "sha=$(git rev-parse --short=12 HEAD)" >> "$GITHUB_OUTPUT"
|
||
|
||
- name: Build rust (musl static)
|
||
run: |
|
||
export PATH="$HOME/.cargo/bin:$PATH"
|
||
cargo build --release --target x86_64-unknown-linux-musl -p "$APP"
|
||
|
||
- name: Run tests
|
||
run: |
|
||
export PATH="$HOME/.cargo/bin:$PATH"
|
||
cargo test --release -p "$APP"
|
||
|
||
- name: Build & push image
|
||
env:
|
||
REGISTRY_TOKEN: ${{ secrets.REGISTRY_TOKEN }}
|
||
run: |
|
||
echo "$REGISTRY_TOKEN" | docker login registry.famzheng.me -u mochi --password-stdin
|
||
docker build --no-cache -f "apps/$APP/Dockerfile" -t "$IMAGE:${{ steps.tag.outputs.sha }}" .
|
||
docker push "$IMAGE:${{ steps.tag.outputs.sha }}"
|
||
|
||
- name: Initialize K8s resources
|
||
run: kubectl apply -f "apps/$APP/k8s/all.yaml"
|
||
|
||
- name: Roll out to k3s
|
||
run: |
|
||
kubectl -n llm-proxy set image "deploy/$APP" "$APP=$IMAGE:${{ steps.tag.outputs.sha }}"
|
||
kubectl -n llm-proxy rollout status "deploy/$APP" --timeout=120s
|