45 lines
1.2 KiB
Markdown
45 lines
1.2 KiB
Markdown
# Qlockify Deployment
|
|
|
|
This repository is the deployment layer only.
|
|
|
|
Docker builds read from the local `./backend` and `./frontend` directories inside this repository.
|
|
Those directories are expected to contain the backend and frontend application source before you build for deployment.
|
|
|
|
## Local structure
|
|
|
|
The expected deployment layout is:
|
|
|
|
```text
|
|
qlockify-deployment/
|
|
backend/
|
|
frontend/
|
|
nginx/
|
|
postgres/
|
|
docker-compose.yml
|
|
```
|
|
|
|
## Deployment flow
|
|
|
|
1. Put your application source into:
|
|
- `./backend`
|
|
- `./frontend`
|
|
2. Configure deployment env files:
|
|
- `./.env`
|
|
- `./backend/.env`
|
|
- `./frontend/.env`
|
|
3. From `qlockify-deployment`, build and start the stack:
|
|
|
|
```powershell
|
|
docker compose up --build
|
|
```
|
|
|
|
The backend container runs database migrations and `collectstatic` on startup, then serves Django with Gunicorn using `config.wsgi:application`.
|
|
|
|
## SSE Notifications
|
|
|
|
Notifications now use Server-Sent Events at `/api/notifications/stream/`.
|
|
|
|
- Nginx disables buffering and cacheing for the SSE endpoint.
|
|
- The current Gunicorn setup can serve SSE for MVP traffic, but each live stream consumes a worker while connected.
|
|
- If notification concurrency grows, move the SSE endpoint to an async worker class or a dedicated ASGI process.
|