Working out what broke.
WHAT TO ESTABLISH FIRST
Whether the failure is real.
WHAT THE POSSIBILITIES ARE
A genuine defect A test that is wrong An environment problem Flakiness A legitimate behaviour change
WHAT TO CHECK
Whether it fails consistently Whether it fails locally Whether other tests fail similarly What changed recently
WHY OTHER FAILURES MATTER
Many tests failing together usually means the environment, not the code.
WHAT TO DO WITH INTERMITTENT FAILURE
Run it repeatedly, in isolation and under load.
WHAT THAT DISTINGUISHES
Flakiness from a load-dependent defect.
WHAT TO EXAMINE
The screenshot or recording The logs from the application, not only the test The request and response, where applicable
WHY APPLICATION LOGS SPECIFICALLY
The test sees symptoms; the application logs the cause.
WHAT TO ADD WHEN DIAGNOSIS IS DIFFICULT
More detail at the point of failure.
WHAT TO NEVER DO
Adjust the test until it passes without understanding why it failed.
WHY
It removes the detection without removing the problem.
WHAT TO RECORD
The cause, once found, on the test or the defect.
