How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyes
[Frontend conference 2019, LT] Automated Test for Frontend
1. Automatic Test for Frontend Side
November 2th, 2019
Frontend Conference 2019 LT Session
Woohyeok Aaron Kim (@woosyume)
Rakuten, Inc.
2. 2
Woohyeok Kim
# Certified Scrum Master
# Speaker at SeleniumConf2019
# Favorite Lang : Korean, Java
Leading E2E Test Automation at Rakuten Ranking
Commit to ‘NO maintenance cost’
Head for integration with our container
Organizing Kansai branch of JSUC(Selenium User Community)
Organize E2E Test Automation Day 2019
Plan to hold study community periodically
Having experience of sessions
Agile Japan 2019
DevRel / Japan Conference 2019
KanJava
Etc.
8. 8
Data Integrity Check
Ranking Result
- Rank
- Item Name
- Shop Name
- Price
- Review#
- Particular condition
Update DateGenre Name
ランキングデータ整合性の確認
1)
1) Ranking Top, 2019/04/03, https://ranking.rakuten.co.jp
9. 9
Image Comparison
Increase test accuracy by checking if image is broken or not.
Calculates difference with the previous execution
Before After
TEST FAILS
画像比較による文字化け・画面崩れの検知
1)
1) Ranking Event, 2019/04/03, https://ranking.rakuten.co.jp/event/anpanman/
10. 10
How to do
DOM
1) Ranking Top, 2019/03/07, https://ranking.rakuten.co.jp
searchBar
contentsTitle
//*[@id”rankingImage”]/ul
id / class / xpath
2분
또 하나는 화면 비교를 통한
1위부터 80위까지 표시
만약 밑에서 모지바케가 발생한 경우 이를 찾아내기가 쉽지 않다.
1프로라도 오차가 발생할 시
%는 허용범위
일상적으로 엔지니어가 발견하기 어려운 것들을 대신 발견할 수 있도록.
엔지니어의 시력, 노안, 컨디션
Based on Appium and test.ai
I tried to indicate DOM by using some trained keywords.
The trained keywords can understand what part is a cart and what image can be used as cart.
If you wanna get an element of ‘KaimonoKago’, you can just set ‘cart’ keyword which is trained beforehand.
단순히 테스트를 기계의 힘을 빌려 한다는 것 이상의 의미를 가짐
소화할 수 있는 양에는 한계가 있다.
보틀넥