Here are a two interesting facts about Azure Backup Server Agent's scheduling:
(Note that this is also applicable to Data Protection Manager)
Time zones handling
During the setup of a Protection Group, you define the time at which the backup should start, but what if you have agents running on server located in different time zones?
The scheduling of your backups will actually follow the local agent's time, not the backup server itself.
For example, if you have a protection group with say 5 servers in there, configured to start backing up at 8pm, and 4 of the servers are in the same time zone as the backup server, while one is one hour behind, you'll notice that only 4 jobs will start at 8pm for those servers located in the same time zone. The fifth server will start at 9pm (which in this example is 8pm agent time).
This makes sense as the recovery point for any workload should be viewed as the local time (not the backup server time).
Ever wonder how MABS schedule and starts jobs?
Well, it kind of doesn't as the scheduling of jobs is actually handled by the SQL server used with MABS.
If you look on your SQL server, you'll see that all the jobs scheduling is under the SQL agent component, which is very handy to know when you want to troubleshoot scheduling failures.
Now based on the resources available on the server itself, this partly why it is good to stagger your jobs across multiple Protection Groups so they can start (and be scheduled) at different times, or to use the "offset" functionality.