Python service for Kubernetes on Wodby

Build and run Python applications on Kubernetes with Wodby.

This repository defines the Wodby service manifests and operational configuration for Python.

• Browse Wodby services
• Wodby service documentation
• Service manifest reference

Start with a boilerplate

Use one of the boilerplates exposed by this service to start with compatible build configuration and Wodby CI:

• Python boilerplate

Wodby stacks using this service

• Python application stack

Service overview

Property	Manifest configuration
Service name	python
Type	Application service
Versions	3.14 by default; also available: 3.13, 3.12, 3.11, 3.10
Workloads	main (Deployment, primary)
Containers	python using wodby/python, build target
Endpoints	python: HTTP 8080 (main)
Service links	DBMS (db), optional; Mail Transfer Agent (sendmail), optional; Redis (redis), optional
Application build	Git source connection enabled; Dockerfile: Dockerfile; boilerplates: Python boilerplate
Helm	chart oci://registry-1.docker.io/wodby/python; version 0.2.2
Configuration and operations	1 integration slots

Use this service

Use this service through Python application stack, or reference python from a custom Wodby stack.

A service is a reusable component and does not deploy by itself. The stack defines its links, settings, versions, resources, and relationship to the rest of the application.

Maintain a custom version

1. Fork this repository.
2. Edit the service manifest and referenced files.
3. Import the repository as a Git-backed service.
4. Reference the service from a stack manifest.

Keep service, workload, container, endpoint, link, volume, config, and derivative names stable unless dependent stacks and app-level overrides are updated at the same time.

Validate the manifests with:

wodby service validate-manifest service.yml --org <org-id>

See the service manifest reference and the managed services index.

Development workspaces

workspace-python prepare runs uv sync --locked when uv.lock exists, or installs requirements.txt into a virtual environment. workspace-python start activates it and runs Gunicorn with its polling reloader against GUNICORN_APP. Override WORKSPACE_PYTHON_COMMAND for another server; HOST and PORT default to 0.0.0.0 and 8080. Uvicorn commands receive WATCHFILES_FORCE_POLLING=true unless explicitly configured otherwise. Custom commands must implement their own reload behavior. Dependency changes require preparation again.

Requires a runtime image declaring workspace contract version 1. Ordinary and development option tags must use matching revisions.