Jason Harmon

I would expect to see the following events at the debug level: * cycle start * spare decision * reap decision (at info level) * cycle end I can see...

I think for backwards compatibility, it's OK to enable all rules by default. If a rule is not configured, or there are no decorated pods, then it does nothing. If...

> @JordanSussman Would you propose that the pod defined label(s) should always override the pod-reaper defined behavior? Feels like a minimum the default behavior should be to ignore pod defined...

> Should a reaper honor a rule specification on a pod when it doesn't have a configuration for that rule? My initial idea was “no” to keep things simple. However...

In general, I like the direction of the refactor. The rules returning a named result instead of a boolean, makes things much clearer what the expected behavior is without having...

@brianberzins Have you made any more progress on annotations? I just ran into a situation where they would be very useful. I'm willing to lend a hand getting this implemented.

I had an immediate need, so I went ahead and implemented this in #64. I'm open to changes.

I like where #45 was going. I had hoped to start with it, but I had an immediate need, so I started with `master`. I'd be happy to refactor this....

The `HONOR_ANNOTATIONS` setting is unnecessary. Any rule loaded with a default will use the default if the pod is not annotated. In other words, behavior will be exactly the same...

Even though annotations are always checked, I would argue that they are still opt-in. 1. Pod owner opts in by annotation a pod. 2. pod-reaper owner opts in by specifying...