Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Scale tick tests are flaky #21642

Closed
markov00 opened this issue Aug 3, 2018 · 11 comments · Fixed by #71452
Closed

Scale tick tests are flaky #21642

markov00 opened this issue Aug 3, 2018 · 11 comments · Fixed by #71452
Assignees
Labels
failed-test A test failure on a tracked branch, potentially flaky-test Feature:Visualizations Generic visualization features (in case no more specific feature label is available) skipped-test Team:Visualizations Visualization editors, elastic-charts and infrastructure

Comments

@markov00
Copy link
Member

markov00 commented Aug 3, 2018

UPDATE LeeDr - please list the completed name of at least one test so we can find the issues;
"visualize app line charts switch between Y axis scale types should show ticks on selecting log scale"

These tests are currently flaky because the window sizes between local test environment and the CI environment differs.

The tests are meant to checks for correct displaying of vertical axis tick labels. The placement and their values depends on the height of the chart, and in visualize editor the height is determined by the height of the window.

I've two different situation on local machine (osx) and on CI:

CI (report a screenshot size of 1280 × 695) with 20 ticks ( 2,3,4,6,8 and their 10x multiple)
local (report a screenshot size of 2560 × 1372 2x dpi because on retina) 19 ticks (2,3,5,7,10 and their 10x multiples minus 10.000)
So on local machine is taken ad 2x density pixel. the different is on height.
The default configured height in leadfoot is 1280 x 800 and

CI have a top chrome space of 105 pixels
local have a top chrome space of 228 pixel at 2x, that means 114 pixels
A difference of 9 pixels in height between the two environments.

CI screenshot
ci

Local screenshot
local

@markov00 markov00 changed the title Skip scale tick tests because of flaky Scale tick tests are flaky Aug 3, 2018
@markov00 markov00 added Feature:Visualizations Generic visualization features (in case no more specific feature label is available) v7.0.0 v6.4.0 flaky-failing-test labels Aug 3, 2018
@markov00 markov00 self-assigned this Aug 3, 2018
@markov00
Copy link
Member Author

markov00 commented Aug 3, 2018

@silne30 Do you know if it’s possible to set the window innerHeight instead of the outerHeight when setting the windowsSize with leadfoot?

@markov00
Copy link
Member Author

markov00 commented Aug 6, 2018

@liza-mae What does "done" means on https://github.com/elastic/kibana/projects/6#card-11860491 ?when the test is skipped or when the flaky test became a stable test? I've already skipped the test on this PR #21641 but I want to keep open this issue so I can find a better solution later

@liza-mae
Copy link
Contributor

liza-mae commented Aug 6, 2018

@markov00 good point, 'Done' should mean the flaky test became stable test, for skipped test we can leave in 'ToDo' or we can create another column 'Skipped' - a test can stay in that state until it is re-enabled, how does that sound?

@markov00
Copy link
Member Author

markov00 commented Aug 6, 2018

@liza-mae I think adding a column skipped can help us to identify the status of an issue. I can move this issue to the skipped, and than move to done when finally the test became stable.

@liza-mae
Copy link
Contributor

liza-mae commented Aug 6, 2018

@markov00 - Thanks! I agree, I have added a skipped column.

@cuff-links
Copy link
Contributor

@markov00 There is no indication of a way to specify but Leadfoot has been known to welcome updates and PRs. https://theintern.io/leadfoot/module-leadfoot_Command.html#setWindowSize
Perhaps a PR to where you can set a flag for innerHeight and outerHeight would be good to do.

@chrisdavies
Copy link
Contributor

In reporting, instead of setting the window height, we set the containing element height. That way, no matter the window size, we get a consistent height in reports. It's hacky, but it'd make these tests predictable.

@rayafratkina
Copy link
Contributor

Removing the version labels: we should not be assigning version until the PR is merged.

@liza-mae liza-mae added failed-test A test failure on a tracked branch, potentially flaky-test and removed flaky-failing-test labels Nov 1, 2018
@stacey-gammon
Copy link
Contributor

@markov00 - you are assigned, are you still working on this, or should this be closed? I think those tests are still being skipped...

@markov00
Copy link
Member Author

markov00 commented Jan 8, 2019

The tests are skipped. There isn't a valid solution to that problem, neither leadfoot or webdriver have a way to reliably set the page height as described on that issue. I've talked to @dmlemeshko today, he will check if it's possible, at least in chrome.

@markov00
Copy link
Member Author

hey @dmlemeshko any news on this topic?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
failed-test A test failure on a tracked branch, potentially flaky-test Feature:Visualizations Generic visualization features (in case no more specific feature label is available) skipped-test Team:Visualizations Visualization editors, elastic-charts and infrastructure
Projects
None yet
Development

Successfully merging a pull request may close this issue.

9 participants