Terminal session process very long

Hello tout le monde,

I’ve got a process with a terminal session under REFramework. 1 000 transaction Item, the job lasts 2h00.

However I only have one transaction to execute where I enter 5 values, validate and check if error message or not.

When I export the log file of my job, I have 10,000 lines (REframework basic)
Can writing a message significantly increase the execution time of the process?

I played with the delay from 300 to 50 and I gained 15 minutes

What can be the areas for improvement for my process?

Thanks

Arnaud

Hey @Arnaud_G Yes excessive logging absolutely impacts performance, especially in REFramework.

Even though logging feels lightweight, over thousands of iterations or very verbose logs, it adds disk I/O, Orchestrator API calls (if logging to Orchestrator), and internal memory usage.

  • Go to Framework\GetTransactionData.xaml, Process.xaml, etc.
  • Comment out or reduce Log Message activities that aren’t essential.
  • Avoid logging at Trace or Verbose levels unless debugging.

cheers

Hello @Arnaud_G

For better performance of the terminal activities you can try the following:

  • For SendControlKey activities, change the TimeoutMs value from 1000ms to 300ms
  • For all other terminal activities, change the TimeoutMs value from 300ms to 0ms

Also, what is the provider you are using for the terminal session?

Regards
Bogdan

For such a simple process, REFramework is huge overkill. You’d be better off writing a slimmed-down project from scratch.

i tried your solution but forgot a few log message arghh . I catch them and run another process with Bogdan’s solution too

Now 1 h for 1000 transaction Items, but too long again.
Thanks for you advice, i’m trying now without REFramwork

I’m not sure what you’re expecting. 1000 per hour is fast. Are you using a dispatcher/performer model and working out of a queue for the performer? Then you can run multiple concurrent jobs to complete everything in less time.

I use a queue with 1 000 transaction Item integrated by another service, a trigger on the queue triggers my process
I cannot use several jobs simultaneously because my connection ID must be unique at a given time, so I must proceed in a linear manner.

Our mainframe works the same, so we have multiple service accounts and set them per robot in the credentials Asset. That way we can run multiple jobs at once.

yes, something to think about, because we are starting on the UI path

Thanks for all

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.