It depends on your license as how many Qlik servers you can use in one multi node. Default is 7, 1 central and rest is on your business use case. each node could have one engine. there could be a node which hasn't any engine lets say proxy service node.
Does the number after 'PRODUCTLEVEL' tell how many nodes we can use in one cluster? For example, we have PRODUCTLEVEL;50;; in LEF, does it mean we can have max 50 nodes in 1 cluster?
Our current set up is 6-node cluster environment with 2019June version. In our experience with QlikView, we had couple of resource-sucking documents which caused QVS nodes not responding at all. All customers in the same node suffers. Therefore, when we design Qlik Sense architecture, one idea is to isolate customer environments. For example: each customer has own dedicated engine node for rendering. Or maybe each customer has own dedicated Qlik Sense environment.
That's why I'm asking how many dedicated engine nodes we can have in each cluster.
We have tried performance tuning. But if an app has all the historical changes, it is very difficult to make improvement without breaking anything. Therefore probably some kind of 'isolated' environment might be helpful, if it is feasible.