modules/nodes-descheduler-profiles: Clarify RemovePodsHavingTooManyRestarts as sum

Pods don't restart, although they can be deleted and replaced with new
pods.  Containers within pods restart.  Previously it was unclear
whether the threshold was "one of my containers has more than 100
restarts" or "summing over all of my containers, there have been more
than 100 restarts".  This commit makes it clear that it's the latter,
based on the backing code [1].  For example, a pod with three
containers, each with 50 restarts, would be eligible for descheduling.

I'm also not super exited about the "removes pods" language, since as
the docs point out elsewhere, some pods are not descheduled (e.g. pods
in the openshift-* and kube-system namespaces).  I'd prefer language
like "Considers pods whose containers...".  But that's orthogonal
enough that I'm leaving it out for this commit.

[1]: https://github.com/kubernetes-sigs/descheduler/blob/bfd5feaf60a202cecafcf2d30cf171a138c04885/pkg/descheduler/strategies/toomanyrestarts.go#L110-L122

Author: wking

modules/nodes-descheduler-profiles.adoc

@@ -28,9 +28,9 @@ It enables the following strategies:
 +
 It enables the following strategies:
 +
-* `RemovePodsHavingTooManyRestarts`: removes pods that have been restarted too many times.
+* `RemovePodsHavingTooManyRestarts`: removes pods whose containers have been restarted too many times.
 +
-Pods that are restarted more than 100 times are removed. This includes restarts for Init Containers.
+Pods where the sum of restarts over all containers (including Init Containers) is more than 100.
 
 * `LowNodeUtilization`: finds nodes that are underutilized and evicts pods, if possible, from overutilized nodes in the hope that recreation of evicted pods will be scheduled on these underutilized nodes.
 +