Have you encountered flaky tests? What practices are there to combat them?

Yes, I have encountered flaky tests — tests that sometimes pass and sometimes fail without changes in the code. The main causes of flaky tests are: asynchrony, dependencies on external services, environment issues, or timing problems.

Practices to deal with them:

• Test isolation: minimize dependencies on external systems, use mocks and stubs.
• Eliminating race conditions and waits: apply explicit waits and synchronization.
• Re-running tests: execute flaky tests multiple times to confirm results.
• Analysis and fixing: identify causes of instability and address them.
• Use stable test data and environment.

For example, for UI tests, you can use waits for elements to appear instead of fixed delays.