The Deadly Consequences of Flaky Tests: Delete Them Now!

[@surajkaintura7968]

You are absolutely on point saying that the first reaction when the pipeline fails is to just RETRY just to see it fail again. Instead of quickly re-triggering the pipeline, we should take a step back to understand the reason it fails. And addressing the problem will reduce a lot of burden on the team compared to just ignoring or deferring it.

[@gui.ferreira]

Absolutely Otherwise, you risk to make Retrying a habit

[@Akira-xc2zi]

It would be nice if you mentioned why retrying is bad.

[@gui.ferreira]

kzbin.info/www/bejne/rZLQqoWQialmaqssi=AXKgPR26QrJy3F-v&t=100

[@impeRAtoR28161621]

I dont agree even remotely. Having Retry(2) on few important tests that are often integration tests (I like to say that one integration test that touches real database, and real systems is like 50 unit tests) is better than delete them completely. Ok, adressing thiose tests should be made (stronger or more build machines, same changes in code) but until them I will not delete them from source code :) I dont say we should ignore them, I just say that if they are important and cover important part of code they should remain. And someone should found flacky tests (probably will have Retry attribute) on them and try to find solution to fix them

[@gui.ferreira]

In the following sentence, do you agree with me that instead of a "should" it should be a "must"? If so, I think you agree with my point 😜 "someone should found flacky tests ... and try to find solution to fix them"