I am writing to you all mainly because I need some guidance with how Qlik Cloud handles cache and RAM with a large app I have. Additionally, the app itself uses 20GB of RAM, and some of the objects are embedded in a webpage for external users to use. I understand that the first time the app is accessed, it will take some time to load due to the size, however, the second time, third time and so on, the app will load faster as it is already cached in memory.
On the other hand, I also understand that when users have stopped using the app (there are no sessions), it can be evicted from memory. Based on this, does anyone know how often does Qlik clear this cache from apps? Is it always checking if apps are no longer used? Or is it a random event based on something?
Any help and or guidance will be greatly appreciated!
Hi, this may be out of date (and fuzzy memory), as I had a relevant call regarding this over a year ago.
For Large Apps, it to use a separate capacity compared to rest of your tenant. For example your dedicated capacity can be 120gb, but your 20gb app will be shifted dynamically on first session to a 80GB server with 10 CPU. From memory, the concurrent users after will be as you said.
I cannot remember the timeout, but for some reason 2 hours stick to my head right now.
To clarify, in Qlik Cloud each App run on a separate Engine based on the RAM Footprint. Default Apps (based on License 5/10/15 GB) or if larger need Large App Capacity based on the Footprint.
For Scaling the first user will need some more time to open, beside loading into Memory we also need to assign an Engine. Every other user will get quicker as we have an active Engine, only case it can took longer is when we see Engine is reaching Limits as we will spinn up a second Engine for others.
When the last user session ended we shut down the Engine 2 minutes later.
