At the moment when we run a test suite. Katalon takes a lot of time just because of logging each step for test suits reports. It adds so much time for a testsuite to run and complete. Just because katalon is logging all the unnecessary steps which we don’t need sometime. For example, we have a testsuite with 172 testcases it takes 18 hours to run. This can be reduced easily if we have option to remove these steps from report and only add helpful info inside a report.
You are right.
See my previous post:
You can close the Log Viewer tab completely. Then, I believe, your test will run in 2 hours.
Detaching Log Viewer and then closing it did the trick. Thanks
Please notify your developers that the Log Viewer’s performance problem is getting well-known. Katalon Studio runs too slow due to the Log Viewer. It could ruin the reputation of the whole brand of Katalon Studio.
2 hours looks too long for me. I want any kind of tests to finish in 10 minutes.
If I were to try to make my tests to run far quicker, I would try running my tests in AWS Lambda.
I do not know if Katalon works in AWS Lambda. If I find Katalon does not work in AWS Lambda, I would regard Katalon is not scalable enough.
In that case, I would switch to Selenium4 in Java. I am sure Selenium4 would work in AWS Lambda. I will divide a long-runing test suite into multiple/smaller test suites each of which finish in 10 minutes. And I will run them with sufficient number of AWS Lambda instances in prarallel.
I would deploy as many Lambda instances as needed: 2, 5, 10, 20, or more. The goodness of AWS is that it charges us by Pay-as-you-go; it does not requires us to buy maximum number annual software licenses.
There is an official tutorial by AWS how to do it:
- Serverless UI testing using Selenium, AWS Lambda, AWS Fargate, and AWS Developer Tools | AWS DevOps Blog
Some people have already worked out with Playwright in Lambda. See GitHub - JupiterOne/playwright-aws-lambda: Support for running Microsoft's Playwright on AWS Lambda and Google Cloud Functions, which proves that we can scale our test using AWS Lambda and Google Cloud Functions.
Actually, the only challange is to properly write the tests to be easily parallelized, with or without Katalon.
For the rest, any ochestrator + whatever CI tool will do it if it is used properly.
Having this in mind, Playwright may be a better option compared with selenium.
and: Tim Deschryver
For a python-top-of-selenium aproach, robot-framework + pabot may be also a decent solution.
Parallelization can be achieved either by running the full suite on the same node but with pure multi-processing (requires a machine with good hardware resources) or by throwing the jobs/threads on independent executors, containerized or not … or whatever else, there are plenty solutions available.
But in any case, a good design of the test suites is a must.