Documentation on internals of timer handling

Hello
Is there some documentations on how the timer are implemented, both for Cassandra and PostgreSQL?
I’m trying to understand the key qualities of that part of temporal, especially interested to understand how timers are honored (sub second precision? what happens in case of burst? which part of the cluster must be scaled, which metrics can be monitored? Is it somehow limited by number of shards?).
Thank you for any pointer