I noticed that by starting a reload qlik dropped some parts from the RAM but not all of it - maybe there are also some dependencies to various settings like compressing, section access and further more how qlik handled it. Therefore it could be possible that if you removed all data before starting the reload that you are starting it with a lesser RAM consumption. If you are quite close to your RAM limit it might be this amount if you goes over the limit or not.
I think if I want to do such thing I would use a preceeding task which opened the application without data and stored it again before triggering the real reload. But I don't think that you will get a stable solution with it if you are so near of the limits. Therefore I suggest to look for any optimizations within the application and often there is much potential of it.
You are certainly right, there is surely lots of potential for optimization in our apps. Some of them are really really big, loading data from like 5years back - who cares for that, nobody knows ... and apart from that, too, there is surely a lot one can do. That's not my main Task, so i can only do it bit by bit, but I'm confident. I already did some optimizing here and there and I'm curious to see how much is possible.
Doubling the number of Tasks we have in the QEMC by using that preceding Task you suggest is not an Option since we already have a large number of Tasks and People are losing oversight which is also one of the issues I want to address whenever I can ... so I end up with the possibility of building an INCLUDE file to Loop through all tables, if any, in an app and dropping them one by one right at the outset.
<=> Unless I could somehow integrate that into the one existing Task by just adding one more Trigger, something like 5min earlier, and making that somehow hit an EXIT right at the outset, thereby emptying the app of all data before then reloading once more. Sounds Kind of tricky - can't wait to try ...
Generally, not everything we have set as a Standard here makes sense ;-) I'm going to try to optimize and standardize what I can and question the Standards wherever necessary ...