We don’t have our devs on call at all. Infra / platform ops are and I think they get 750€ per on-call week (not more than one week out of four) which includes two calls or two hours of call duration whichever is reached first.
After that it’s another 70€ per call or started hour and it’s the same if an expert who is not on call is asked to help out with an issue reported to on-call (but they may not answer / decline as there’s never an expectation to be “soft on-call”)
Overall that’s an okay deal and some sorely needed extra money for the ops guys and gals. But all the same I’m happy that my devs don’t need to plan their lives around an on-call schedule.
Edit: Ah sorry, didn’t even answer all the questions in OP…
We’re in Germany and there is a cooldown time after you fielded an emergency on-call report (which is outside of regular working hours by definition) which is either 8 or 10 hours (not entirely sure since my team doesn’t do on-call as previously stated) before you are allowed to start your regular work time for the following day.
Not sure how they tally up working hours for payroll but if you wake up to a call at 3am then certainly no one expects you to be online again at 8am. If you get a call at 10pm however then you get to start working normally the next day. (unless that issue took forever to troubleshoot ofc)
On-call rotations are one entire week per person who participates (which is not mandatory) and the participants per pool must be at least four - which is why they are pooling web admins, DBAs and other ops folk together.
That seems to work okay even though every so often more specialized know-how is required than the current on-call tech possesses for the topic at hand and then they request extraordinary assistance as described above.
IBM sure does suck donkey balls but I really don’t think that particular thing is their fault.