I’ve seen the same at several different companies. There seems to be just too much that can go wrong during an e2e test like this: network failures, slow CPU ca

I’ve seen the same at several different companies. There seems to be just too much that can go wrong during an e2e test like this: network failures, slow CPU causing timeouts, etc. At one company, the entire suite of 700 tests was re-run 3 times nightly, taking 5+ hours, and some tests still failed randomly. I have found that the majority of errors can be caught by unit and integration tests on a single service, and that the extra burden of maintaining Cypress just isn’t worth the small amount of coverage.