https://github.com/schrodingdong/zeta-functions
Local serverless implementation, because if it is possible, we can.
https://github.com/schrodingdong/zeta-functions
docker fastapi python serverless-functions
Last synced: 2 months ago
JSON representation
Local serverless implementation, because if it is possible, we can.
- Host: GitHub
- URL: https://github.com/schrodingdong/zeta-functions
- Owner: Schrodingdong
- Created: 2024-11-02T11:53:07.000Z (8 months ago)
- Default Branch: main
- Last Pushed: 2024-11-22T00:18:24.000Z (7 months ago)
- Last Synced: 2025-04-04T11:23:19.439Z (3 months ago)
- Topics: docker, fastapi, python, serverless-functions
- Language: Python
- Homepage:
- Size: 126 KB
- Stars: 4
- Watchers: 1
- Forks: 0
- Open Issues: 9
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Zeta - The local open-source serverless project
This project will try to mimic serverless computing localy on your device.
## Why would you need it ?
idk its cool tho.
## Quickstart
> Right now, the project is tested on linux / WSL. Make sure that the project is set up in the same host as the one docker is running in.
> For WSL make sure to check the `Use the WSL 2 based engine` and `Enable integration with my default WSL distro`.
- Pull the project
- Create a venv. Activate it and install the requirements
- Run this command
```bash
# Export the DOCKER_SOCKET variable in this shell instance
# For global use (not recommended because of sudo), add it to your .bashrc file or similar
export DOCKER_SOCKET=$(sudo find / -name docker.sock | grep docker.sock)
```
- Build the runner images
```bash
# Python Base runner
docker build -t python-base-runner:latest ./src/runner_images/python_base_runner
```
- Run the fastapi host application, which is a docker-proxy
```bash
fastapi dev ./src/docker_proxy/main.py
```
- Create a Zeta function `POST localhost:8000/zeta/create/`
- payload is a python file, with a `main_handler` as its entrypoint:
```python
def do_some_computation():
# ...
def main_handler(params):
# Logic ...
return { ... }
```
> Technical Note: Using the same python file for multiple function deployment will result in multiple runner images generated, with the same imageID.
> That is because they are using the same layers. This shouldn't impact the app execution, but the more you know ;)
- Run the function `localhost:8000/zeta/run/`
- payload should be the same as used for the handler
## To use the CLI
Here are the commands supported:
```
VERSION ===========================
zeta version
CREATE ============================
zeta create
- (Re)Create / (Re)Deploy the zeta, and returns its url for the user
DELETE ============================
zeta delete
- Deletes the zeta
NAME LIST =========================
zeta list
zeta ls
- List zeta names (ONLY)
INFO ==============================
zeta ps
- List all zeta metadata
zeta ps
- Returns zeta metadata for the specified zeta
```
## Supported Languages
- [x] Python
- [ ] Java
## Requirements
- The User should define functions in a supported language, which will be defined as a "Zeta Function"
- Creating a Zeta function will build an image following this name convention:
- `-runner-image-`
- A Zeta function should instanciate a container to execute the function
- The container lingers for 5 minute before stopping if no activity is detected
- Concurrency: Each user will have their containers separated from the other ones
- Auto-scalability: If there is to much load on a container, make sure to scale it horizontally