I'm looking for a way to restart all the pods of my service. They should restart one by one so the service is always available. The restart should happen when a Python script from a different service is done.

I'm doing this because on the pods I want to restart there is a Gunicorn-server running which needs to reload some data. That only works when the server gets restarted.

The gunicorn service gets started in a Dockerfile:

CMD gunicorn -c gunicorn.conf.py -b :$PORT --preload app:app

But I'm guessing this is not too relevant.

I imagine the solution to be some kind of kubectl command that I can run in the Python script or a hint for a kubectl endpoint, that I couldn't find.