Rudder

Following #4766, check-rudder-agent now takes into account the frequency of agent execution.

We decided to double the run interval so we could handle the splay time correctly and that we are sure that we haven't miss any execution.

But this case only occurs if the splaytime is very close to the frequency, otherwise the check is made really late and could have been done before:

if the frequency is every 2 hours, and the splaytime 5 minutes, if the agent  has ran a 0:00, next agent run can occurs between 2:00 and 2:05
If the lock happens here, it will only be detected at ~ 4:00, so have to wait at least 1h55 before fixing the issue.

The check should look every "run Interval + splaytime + 5 minutes" (to ensure that the agent has finished to run), so you have to wait only 5 minutes after splytime to fix tcdb issues.

To do that we need a file where the splaytime is stored, like the run interval.