Rudder

I honestly have NO IDEA on how to fix this. If the agent execution is not often engouh (6h interval for agent execution against a 2h time schedule) it will completly miss the schedule :/

Is there anything in CFEngine 3.6 that could help refactor that (quite hard) technique in a simpler way - and allows to remove that issue ?

There is also the question of the semantic we want to have for the missing runs: do we run them as soon as we saw that one is missing, or do we way for the next interval ? And of course, the second option must not lead to a "never run the job" if we are in the case matching the current issue (i.e, we have to be able to decide between "as close as we could have been for the job datetime - so ok, say it the same instant" from "we missed a job schedule, because perhaps the machine was down, what do we do with missed jobs ?")