Updated README with steps to fix current issue

SWAOS
2026-01-29 21:25:48 +00:00
parent bd16f55f2e
commit 1c5300a6f4
1 changed files with 146 additions and 42 deletions
--- a/README.md
+++ b/README.md
@@ -1,43 +1,147 @@
-Here is dummy ai explanation for you my friend
+ ### QUICK EXPLANATION

-## Stack
-* Backend: Node + Fastify + Prisma (Postgres)
-* Frontend: React + Vite
-* Scheduler: separate worker container (cron tasks)
-* Reverse proxy (prod)
-* Default: Nginx (skymoneybudget.com.conf)
-* Alternative: Caddy (Caddyfile.prod)
-* API is under /api path
-* Web is static from /var/www/skymoney/dist
-* HTTPS + HSTS already configured in the prod config
- ### Required env vars (production)
-* DATABASE_URL
-* JWT_SECRET (32+ chars)
-* COOKIE_SECRET (32+ chars)
-* CORS_ORIGIN (set to https://skymoneybudget.com)
-* Optional: COOKIE_DOMAIN=skymoneybudget.com
-* Cookies + Security
-* HttpOnly + Secure in prod
-* CSRF protection required for non‑GET
-### Other Features
-* /auth/logout requires CSRF now
-* Mutation endpoints rate‑limited
-* Cron jobs
-* rollover worker (6 AM user time)
-* auto‑payment worker (9 AM user time)
-* Jobs run every 15 minutes and only process users whose local time has reached the threshold
-* DB / Prisma
-* Postgres required
-* Prisma migrations must be applied before running
-* The app blocks default secrets in prod
-* Build + deploy
-* web/ builds to web/dist
-* API runs on port 8080 internally (exposed as 8081 in compose)
-* Nginx proxies /api to 127.0.0.1:8081
-* Logs
-* Prod logs are limited (PII reduced)
-* Key logging: job success/failure counts
-* Backups
-* Scripts exist: backup.sh, restore.sh
-* Restore requires admin DB privileges (created DB)
-* Optional but recommended to test once (tested, backup worked restore failed)
+In order for the system to properly function with all background workers, the deployment needs to be transferred to docker compose. This file explains the proccess. I do not have access to the VPS so I figured I would create a guide. 
+
+- The reason the session is not working currently is because the API runs within docker compose and all the workers can not function (auth for example). 
+ 
+ 
+ # Deploy via Docker (Gitea Actions)
+
+This document lists the changes needed to switch your deployment to **Docker Compose** so all services (api, web, scheduler, postgres, etc.) start automatically on deploy.
+
+---
+
+## 1) Replace the deploy workflow steps
+
+Update `.gitea/workflows/deploy.yml` so it:
+
+- Pulls the repo on the VPS (or uses the runner workspace)
+- Builds and starts containers with Docker Compose
+- Runs migrations inside the running API container
+- Reloads Nginx (or Caddy)
+
+**Recommended deploy.yml (Docker-based):**
+
+```yaml
+name: Deploy
+
+on:
+  push:
+    branches: [main]
+
+jobs:
+  deploy:
+    runs-on: vps-host
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Deploy with Docker Compose
+        run: |
+          # Adjust to your deployment directory
+          APP_DIR=/opt/skymoney
+          mkdir -p $APP_DIR
+
+          # Sync repo contents to server folder
+          rsync -a --delete ./ $APP_DIR/
+
+          cd $APP_DIR
+
+          # Build + start all services
+          docker compose up -d --build
+
+          # Run Prisma migrations inside API container
+          docker compose exec -T api npx prisma migrate deploy
+
+      - name: Reload Nginx
+        run: sudo systemctl reload nginx
+```
+
+> If you use Caddy instead of Nginx, replace the reload step accordingly.
+
+---
+
+## 2) Ensure docker-compose.yml is production-safe
+
+Double‑check these:
+
+- The API service exposes the correct port
+- The web service is either served by Nginx (static build) or a container
+- Scheduler/worker services are included
+
+Example pattern:
+
+```yaml
+services:
+  api:
+    build: ./api
+    env_file:
+      - ./.env
+    ports:
+      - "8080:8080"
+
+  scheduler:
+    build: ./api
+    command: node dist/worker/rollover.js
+    env_file:
+      - ./.env
+
+  web:
+    build: ./web
+    # or copy build output to nginx
+```
+
+---
+
+## 3) Update Nginx upstream
+
+Make sure your proxy matches the API port:
+
+```
+location /api/ {
+  proxy_pass http://127.0.0.1:8080/;
+}
+```
+
+---
+
+## 4) Add Docker/Compose to VPS if missing
+
+```bash
+sudo apt update
+sudo apt install -y docker.io docker-compose-plugin
+sudo systemctl enable --now docker
+```
+
+---
+
+## 5) Update environment files
+
+Your `.env` should exist at `/opt/skymoney/.env` and contain:
+
+- DATABASE_URL
+- JWT_SECRET (>=32 chars)
+- COOKIE_SECRET (>=32 chars)
+- CORS_ORIGIN
+- NODE_ENV=production
+
+---
+
+## 6) Verify after deploy
+
+```bash
+docker compose ps
+curl -i http://127.0.0.1:8080/health
+```
+
+---
+
+## Optional: logs
+
+```bash
+docker compose logs -f api
+docker compose logs -f scheduler
+```
+
+---
+
+If you want, I can also adapt the deploy workflow to use **docker compose pull** + **docker compose up** without rebuilding every time, or split API/web deploys.