We just installed QSDA Pro at a client and are experiencing extremely high resource usage very quickly. We were very excited to start leveraging the insights from QSDA Pro, but were immediately halted in our progress.
QSDA Pro is installed on our on-premise reload node, a 4-core VM with 16GB of RAM. This node is typically only used during the night to reload production dashboards, and is generally idle during the day. Idle RAM usage is around 3-4GB.
Initially, we analyzed two fairly large apps, which we were keen to start optimizing. After opening one of the analyses and applying a filter on non-published sheets, a loading icon appears in the analysis interface. Meanwhile, the server’s CPU usage shoots up to around 80-100% and RAM usage trends upwards rapidly, in a spiky fashion, until fully saturated, causing the entire server to grind to a halt:
Stopping the QSDA service from the Control Panel immediately results in normal RAM usage again. This behaviour can be reproduced and the Error Log in the Admin panel does not offer any additional information.
What could be causing this issue? It almost seems like a memory leak of sorts. We’ve found no system requirements in the QSDA Pro documentation, but 16GB on an otherwise idle server would seem sufficient.