From a code style point of view, everything looks good. There are two cases included in the count here, and I wonder if they should be separated:
Triggers not sent because of lack of tokens
Triggers not sent because triggers were paused
A possible reason to separate them is so that a "clean" run in which we never run out of tokens gets an "inhibited" count of 0. As it stands, you could get a nonzero inhibited count from the triggers that would have been issued between "start" and "resume", even if tokens never run out.
Or, easier than separating them would just be to not include case 2, on the thinking that it's a less interesting quantity from the point of view of understanding whether things are working. What do you think?
From a code style point of view, everything looks good. There are two cases included in the count here, and I wonder if they should be separated:
Or, easier than separating them would just be to not include case 2, on the thinking that it's a less interesting quantity from the point of view of understanding whether things are working. What do you think?