I’m using the recorder via the Chrome plug in on a Mac.
The latest version has a couple of problems with execution speed:
If you have the recorder open with the log live scrolling, then as the test suite runs the execution speed decreases dramatically. The bigger the log, the slower the execution. This didn’t seem to be a problem in earlier versions.
To get around this, I now have a different tab open, such as Screenshots or Variables during execution.
Even so, and even if I clear the log, execution time slows over as I run more test suites. Quitting the extension and starting over speeds things back up again.
Whatever is slowing things doesn’t seem to be CPU dependant. I can see my CPU load decrease as the test execution slows down.
I have this same issue. One way I got around it temporarily was to vastly reduce log level. It still slowed down later in the run, but was much more manageable.
Same issue here, will try to reduce logs and hide log tab to avoid degradation.
Did reduce the logs help with the performance? Can you provide more info on the number of test cases, and if possible, provide us with the execution log file to reproduce this issue?
It’s true, the size of the log (real-time log output) affects the execution speed. I did some tests, I used:
Katalon Recorder 188.8.131.52
Google Chrome, Version 92.0.4515.159 (Official Build) (64-bit)
Play speed in KR – maximum fast
Microsoft Windows 8.1 64-bit, Windows Server 2012 R2 Standard
Test 1: the log is not hidden
We measure the approximate time of each iteration of the loop (more precisely, the execution time of a group of commands between timeWhileStart and timeWhileEnd at each iteration). Then we look at the array with the obtained values (in milliseconds). The log is not hidden, extended.
We got an array of values: the time of each “iteration” 01.txt (493 Bytes)
Average execution (“iteration”) time 452.5 ms. The average difference between the N and N-1 values is 5.8 ms. The average difference between the N value and the N-10 value is 60.79 ms.
A similar result will be if we make one test suite, where each “iteration” will be in a separate test case
time_execution_test_02_b.html (100.7 KB)
Test 2: hidden log
The same as last time, but we hid the log (clicked the toggle “arrow”)
As a result of hiding the log, the time of each “iteration” has decreased:
Average execution (“iteration”) time 85 ms. The average difference between the N and N-1 values is 0.04 ms. The average difference between the N value and the N-10 value is 1.2 ms.
Test 3: hidden log and 999 iterations
We have hidden the log, but increased the number of iterations to 999. The execution time of each iteration is growing, but slower.
Average execution time 145.49 ms. Average difference between N value and N-1 is 0.134 ms. The average difference between the N value and the N-10 value is 1.32 ms.
Wow, thank you so much for the detailed report. We’ll investigate the underlying issue and keep you informed!
Thanks so much, was just going to ask the same question - I iterate through about 16 tables, each of around 20 entries and it can take over an hour to complete. Is it possible to disable the log altogether and get greater speed?
Just tested with a minimised log - and WOW 920 rows done in 6min, not 1hr+. Great pick up!!