jeudi 19 avril 2018

How should continuous integration handle failure-causing external dependencies?

I'm currently developing a web application which is tested in TravisCI using Cypress; part of the tests rely on a company-wide search engine which cannot easily be ported into the CI.

Tests occasionally fail due to the search engine going down; should it be acceptable to ignore these failures IFF the external system is down (which could cause a bug to enter the codebase), or should tests be re-run if this error occurs until the external service is available again?

Aucun commentaire:

Enregistrer un commentaire