Rudder

For now, we use an heuristic based on "2 agent run + some more time" before we expire compliance and set everything to "no answer".

We want to be able to let the use pilot that latency by letting him specify a period of time during which compliance will remain valid, even if it's days.

That setting must be by-node, and if possible by-group-overridable by node, because the typical use case is "nodes in the 'laptop' group should have a 7 days compliance validatity before we expire it, while server in group 'important' should be checked as often as possible".

This clearly point to the use of (global, group, node) properties.

So that ticket goal is to make the "run agent" parameterized by a period of validity, and get the period of validity from properties.