Skip to main content
swarajkumard9182
March 13, 2024
Question

Slow Process model performance on load test

  • March 13, 2024
  • 6 replies
  • 0 views

Hi All,

We are working with on-premise version of Appian (23.3) which is hosted on a private cloud. When I run load test on my environment process models are taking more time to complete (ex: 35 mints to complete).we increased process and execution engines also but still having latency. each node is taking around 2- 3 minutes to execute , even a simple is null logic also taking same time.  We have done the health check but could not see any major issues. Any suggestions to debug this issue would be greatly appreciated.

Note : Load test time we are not running any background Process

Thanks in advance.

6 replies

stefanhelzle0001
March 13, 2024

Node completion time, that you observe here, is not equal to node execution time. This is how Appian scales. But a completion time in the minutes range seems a bit steep. I suggest to contact Appian to discuss that.

swarajkumard9182
March 13, 2024

I created a case with Appian they asked me to increase engines and CPU , but they are not ready to give metrics. to satisfy my load how much big system configuration needed ? at the moment we have 8 CPU and 64 GB space 6 analytic and 6 Process execution engines , above is just a 5% of our load. If we want to run our 100% load I dont know which configuration need to use.  

stefanhelzle0001
March 13, 2024

To repeat that: Long node completion time does not mean high system load.

But, I am used to node completions time in the one second range. Several minutes does not feel right and the only way to find the issue, is to discuss that with Appian.

March 13, 2024

you can check is there any other process or work is happening in the environment which is making other resource occupied

swarajkumard9182
March 13, 2024

No Other processes are running at that time.

hishoren
March 19, 2024

If you are not able to find anything on the health check, you need to increase the server resources. Additionally, the resource details you mentioned(8 CPU and 64 GB space 6 analytic and 6 Process execution engines) will not be sufficient for 6 execution pairs during the high load. It will affect the memory and the execution engine will go down if there is so much load.