My implementation was really simple, I mostly only scheduled jobs in it. My CMS was an FCI for local HA and I used dba tools to copy the jobs to a DR CMS so they would be there available in the case I ended up operating out of DR for an extended period that would need to make job changes. It also synchronized my scale out SSIS jobs that could fail over transparently to the new master at DR and the SSIS catalog database was kept in the AAG, replicated to DR.
All the CMS jobs had something in their name to indicate they were a centrally managed job as well as the database and server group they ran against. If it needed to be switched over, we would just need to pull the target servers out of the original master (this is going back a few years, struggling to remember the terminology of functions we never used) and then register the DR server into the new CMS and edit jobs to target the remaining DR SQL server for all the centrally managed jobs.
It would have been a manual pain to reconfigure all those jobs, but our primary datacenter was so robust that it was our assumption anything big enough to get is to failover to DR would cause a very long change freeze and was unlikely to need to make any changes.