Query on QlikSense Reload Task Optimization: Cores, VCPUs, and Threads

[Charvick]

Hi All,

I am writing to seek clarification on the optimal configuration for reloading Qlik Sense apps. According to best practices, it is recommended to set the maximum number of concurrent reloads to the number of cores minus 2. For example, on a system with 16 cores, the ideal setting would be 14 concurrent reloads. Article - https://customerportal.qlik.com/article/Concurrent-Reload-Settings-in-Qlik-Sense-Enterprise

To better understand how to optimize our server configuration, I have the following questions:

• What is the relationship between cores, VCPUs, and threads in Qlik Sense, and how do they impact reload tasks?
• Do I need to increase the number of cores or VCPUs to accommodate more reload tasks, or is one more beneficial than the other?
• What are the benefits and drawbacks of having more cores, VCPUs, or threads in terms of reload speed and overall system performance?
• Specifically, does increasing the number of cores or VCPUs lead to faster reload times, or are there diminishing returns?
• Are there any recommendations for configuring threads, and how do they interact with cores and VCPUs?

I would appreciate it if you could provide a clear explanation or point me to relevant documentation.

Thanks in advance.

[Eduardo_Monteiro]

Hi @Charvick 

Cores/vCPUs are interchangeable in Qlik's eyes — both increase your concurrent reload capacity.

More cores = more concurrent reloads, not necessarily faster individual reloads (unless the reload itself was CPU-bound).

The golden rule is n−2 concurrent reloads per node, leaving 2 cores for OS and other Qlik services.

For large reload volumes, horizontal scaling (more nodes of equal size) is more effective than endlessly stacking cores on one server.

Threads are relevant for proxy/session handling, not for reload performance directly.

Please let me know if further clarification is needed and I'll be glad to help.

Regards