Rudder

In rudder, there is a grace period for when we get the last reports from a node before marking it as down ("no reports").

This grace period is (by implementation) one run period + 5 minutes. So with a frequency of one run every 5 minutes, that lead to a grace periode of about 10 minutes before marking the node missing.

But we are telling user that the grace period is "two runs", which is false, and so people expect 30+ minutes when period is 15min in place of 20+.

It is only in 3.1 as we remove the text in 4.1.