When conducting code review of colleagues' automated tests, what do you pay attention to first? What red flags do you look for?

When conducting code review of automated tests, I pay attention to the following points:

• Scenario coverage: tests should cover both positive and negative cases.
• Reliability: tests should not be flaky (unstable), it is important that they give predictable results.
• Readability and maintainability: test code should be understandable, with a logical structure and comments if necessary.
• Isolation: tests should be independent of each other, not affecting the state of external systems unnecessarily.
• Use of correct assertion methods: checks should be precise and informative.
• Absence of redundant waits and timeouts: so that tests do not slow down the build.

Red flags:

• Hardcoded data that is difficult to change.
• Complex and long tests that are hard to understand.
• Using sleep instead of explicit waits.
• Tests that depend on external services without mocks.
• Lack of cleanup after tests (for example, created data is not deleted).

Example of a bad pattern:

// Using Thread.sleep instead of wait
Thread.sleep(5000);
assertTrue(element.isDisplayed());

It's better to use explicit waits.