Make documentation about clock drift detection more vague to not leak implementation details

[kevincianfarini]

In the future we might consider that resuming coroutines every minute to check for clock drift is too frequent. For example, we might consider delaying the coroutine for half of the remaining duration several times. This would allow us to only resume delayed coroutines a handful of times to check for clock drift versus checking about 1,440 times per day per pulse.

Right now the documentation specifically mentions that coroutines will be resumed every minute to check for drift. We should be more vague about this and allow this to be an implementation detail. Perhaps the following language would suffice:

Coroutines delayed for more than one minute will occasionally be resumed to check for clock drift.