Since the whole step-one and step-two checks require a lot of time, we should think about saving state and resume from the last checkpoint if the script fails or is aborted while executing. We could save results at the end of each check and resume from there.

We should also avoid opening duplicate tickets in pagure, so let's fetch them at the start of step-one and exclude users for which we already have an open ticket (this will also speed up the process by reducing the number of checked users).