When building Automation framework from scratch what needs to be considered?
How do you calculate ROI of your automation? Is there any metrics
Adding questions from Twitter
I started with Ben’s question.
Here is the video.
What type of tests should be covered in the automation suite?
Thanks for your question, here is my answer
Hey Sravika, I’ve answered your question here
Is it good approach in API testing to deserialize the API response into a POJO and comparing the expected object with resulted object or i can use third party tools like Approval Tests to verify entire response body
What is the context where you need to check the entire response?
Usually, consumers use only a subset* of the responses of an API. Which means that the integration contract between one consumer and one producer relies only on a subset of the response.
* The case where this subset is equal to the set itself is only by coincidence. Future changes in the response to fulfill the need of other consumers do not alter the contract between the first consumer and the producer.
New questions from Linkedin and Twitter
Steve Walters asks “What are your thoughts on how AI / ML could be used for automation in testing? Thanks!”
Beth Marshall asks “Refactoring automation code: why, when and how?”
Beth Marshall asks “Is it better to run API testing through a stand-alone tool e.g. Postman, or add these tests through an existing UI framework?”
well thats my second question and feel free to prioritise later to give chances to others.
Do you agree that automation should be 100%? if not then how would you convince manager/client or other stake holders?
I’ve answered your question
different types of testing can be automation
5.data driven testing
My question is:
Answer to Beth’s question: “Refactoring automation code: why, when and how?”
Would you have any thoughts on this question of mine?
This is Petros from MoT Athens.
So my question is this:
We have proposed to a client a specific automation tool (Selenium with Java) but he/she keep asking for a different automation tool that we don’t have so much experience (and in different language too). Client has an external consultant and pushes us for different things that we recommend. Unfortunately my company said ok to the client “because is the client”. So what do you think about that? Should we always agree with the client? My team is frustrated about this
I have automation tests on Jenkis, which is installed ond Windows Server machine, so there’s some issue with the tests for Internet Exproler. When one test fail, some IE driver or IE is “stucked” in Task Manager and and all tests after that one are failing, driver want start.
Do you have any clue how to resolve to issue or are you maybe familiar with this?
Hello @iamsethi, here is my answer
Nice initiative. My question is :
We maintain POJOS in the stand alone API test project instead of using a shared library from the main project.
The reason why I preferred to do so is to track the changes in the models and update them in the test project if the model change was indeed intended (like an additional net). This comes with the cost of time for maintenance.
What is the best practise here?