Project

General

Profile

action #88183

[spike][timeboxed:20h] rancher: Create a simple selenium test for the UI as proof of concept

Added by ilausuch 6 months ago. Updated 6 months ago.

Status:
Resolved
Priority:
Low
Assignee:
Target version:
Start date:
2021-01-25
Due date:
2021-02-08
% Done:

0%

Estimated time:

Description

Observation

And since openQA is widely used as a testing system. This ticket is intended to prove that the integration of rancher tests with openQA is possible, in particular in the case of the UI

Problem

OpenQA can use the standard method of using needles to test the UI (see https://openqa.opensuse.org/group_overview/78), but since rancher UI is a browser application they propose to use more appropriate tools such as selenium. This ticket covers the creation of a simple selenium test for Ranche UI

User story

I would like to promote openQA as a testing tool, and collaterally create the seed for a Rancher UI test with selenium

Suggestion

Use Python as the main language of the test as it may be more familiar to selenium developers and can be run independently of openQA.

Acceptance criteria

History

#1 Updated by ilausuch 6 months ago

  • Subject changed from rancher: Create a simple selenium test for the UI to rancher: Create a simple selenium test for the UI as proof of concept

#2 Updated by ilausuch 6 months ago

  • Description updated (diff)

#3 Updated by okurz 6 months ago

  • Project changed from openQA Project to QA
  • Due date set to 2021-02-08
  • Priority changed from Normal to Low
  • Target version set to Ready

As discussed, thanks for making your work explicit. I agree that this can be a nice "company integration task" but we have to keep our priorities for the existing backlog in mind

#4 Updated by cdywan 6 months ago

Could you add ACs for this ticket, please? Are Selenium tests part of this ticket? Or setting up existing tests within openQA?

#6 Updated by ilausuch 6 months ago

  • Description updated (diff)

#7 Updated by cdywan 6 months ago

Thank you for adding more details!

  • AC1: Create a test using python to check some parts of the Rancher UI

I would recommend not to precribe the implementation here. Instead, you want tests using Python-bindings for Selenium to be a suggestion. And maybe even compare it to Perl bindings in this context to further evaluate the concept.

Can you confirm that this should be a proof of concept? Or what we usually call a [spike][timeboxed: timeframe ]? So I wouldn't probably make adding regular tests an AC here. That feels like it's skipping a couple steps.

For instance, where should the test code live longterm? The proof of concept based on Python-bindings to Selenium is pretty much agnostic to openQA right now. So I'd expect that code to live in an existing or new project to develop those tests.
Or maybe you want to integrate Python/Selenium actively within openQA? In that case I'd be curious what your plans are!

#8 Updated by okurz 6 months ago

  • Status changed from In Progress to New

acceptance criterias should be written from a user perspective. What you wrote are rather "implementation suggestions". Maybe it helps if you look into existing tickets how ACs are phrased?

Also other references I have just found which might be useful:

As we discussed this ticket let's add it back to "New" to visualize that we do not yet consider it READY as according to our https://progress.opensuse.org/projects/qa/wiki#Definition-of-READY-for-new-features . This of course does not mean that your previous effort was in vain, but is just regarding the format of the ticket.

#9 Updated by okurz 6 months ago

ilausuch we talked about this ticket and AFAIR we agreed that you would bring it into shape before investing more effort but I see you progressed with https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/11863 . Could you please look into the comments from cdywan before potentially wasting effort because we do not even know what we aim for? As it was another 3 days and over the due date I suggest by now that you conclude the ticket as an "experiment" so better just note down the results and resolve the ticket.

#10 Updated by ilausuch 6 months ago

Yes, conclude this ticket as experiment

As talked with cdywan this morning the objective was proof of concept and for us it's done in spite of the errors related of the execution of the tests in openQA. Because we have a selenium script that works. The problem is that I am block to run that inside openQA, but it is not part of the testing that of selenium

#11 Updated by okurz 6 months ago

ilausuch wrote:

Yes, conclude this ticket as experiment

Not sure who should be the actor here. So will you resolve this ticket please with a report about your results?

#12 Updated by cdywan 6 months ago

  • Subject changed from rancher: Create a simple selenium test for the UI as proof of concept to [spike][timeboxed:20h] rancher: Create a simple selenium test for the UI as proof of concept
  • Status changed from New to Resolved

I thought ilausuch was going to but may have misunderstood, so I'll update as per my understanding. The request was meant as a proof of concept to explore what's needed and serve as a basis for discussion. Adding working tests is out of scope here.

Also available in: Atom PDF