- Automatic K8S Deployment
- Authenticate to private registry automatically
- Load Deploy secrets during build step and apply to cluster
Move Turso client initialization/binding to shared worker context.
It's currently created in the app.before_server_start hook and added to app.ctx
This may be the culprit of slow startup time (WorkerManager.THRESHOLD = 600) if each worker process (10 per pod on current testing env) has to initialize a connection to Turso
Docker images are automatically built for amd64 and aarch64
Get started with x86
docker pull git.victorbrink.dev/victor/sanic-server:latest
Get started with arm
docker pull git.victorbrink.dev/victor/sanic-arm:latest
Install uv (pip install uv)
uv venv # Create virtual environment
uv run pip install . # Install project dependencies
Running the app can be done with most of your favourite ASGI or WSGI web servers. I am using Sanic's built in server to utilise their Worker Manager
Development server
uv run sanic server --dev --reload
To run in production remove --dev, configure workers and run with --no-access-log to prevent Sanic from writing an access log.
The Python API is slower than other logging options, you can reverse proxy through nginx for access logs and tune nginx for sanic
This will start a Sanic HTTP server with one worker. Sanic provides an integrated REPL shell in dev mode, just press enter in the terminal.
Sanic is built for speed and creates multiple seperate proceses as workers. Configuration Guide
The easiest way to automatically use the highest worker count, is by using the --fast flag after sanic server you can combine it with --dev
Testing out https://sanic.dev/.