Why Flaky test detection and retry in PyTest? - Purpose & Use Cases

The Big Idea

What if your tests could tell the difference between real bugs and random glitches all by themselves?

The Scenario

Imagine running your tests manually every time you make a small change. Sometimes tests pass, sometimes they fail for no clear reason. You spend hours trying to figure out if the problem is real or just a random glitch.

The Problem

Manual testing is slow and tiring. You can easily miss flaky tests that fail randomly. This causes confusion and wastes time fixing problems that do not really exist. It is hard to trust your test results.

The Solution

Flaky test detection and retry automatically rerun tests that fail randomly. This helps identify if a failure is consistent or just a one-time glitch. It saves time and gives you confidence in your test results.

Before vs After

✗ Before

def test_feature():
    result = feature()
    assert result == expected

# If it fails, you rerun manually many times

✓ After

@pytest.mark.flaky(reruns=3)
def test_feature():
    result = feature()
    assert result == expected

What It Enables

You can trust your tests to catch real problems and ignore random failures, making your development faster and less stressful.

Real Life Example

In a team project, flaky tests caused confusion and delays. Using flaky test detection and retry helped the team focus on real bugs and deliver features on time.

Key Takeaways

Manual test reruns waste time and cause confusion.

Flaky test detection automatically retries tests to find real failures.

This improves trust in tests and speeds up development.

Practice

1. What is the main purpose of marking a test as flaky in pytest using @pytest.mark.flaky(reruns=N)?

easy

A. To skip the test permanently

B. To automatically retry the test N times if it fails

C. To mark the test as slow and run it last

D. To run the test only once without retries

Solution

1. Step 1: Understand the flaky test concept

   Flaky tests fail randomly due to timing or environment issues, so retries help reduce false failures.

2. Step 2: Analyze the effect of @pytest.mark.flaky(reruns=N)

   This decorator tells pytest to rerun the test up to N times if it fails, increasing chances of passing despite flakiness.

3. Final Answer:

   To automatically retry the test N times if it fails -> Option B

4. Quick Check:

   Flaky test retry = reruns N times [OK]

Hint: Retries mean automatic rerun on failure [OK]

Common Mistakes:

• Confusing flaky with skipped tests
• Thinking flaky marks slow tests
• Assuming flaky disables retries

2. Which of the following is the correct syntax to retry a flaky test 3 times in pytest?

easy

A. @pytest.flaky(tries=3)

B. @pytest.retry(3)

C. @pytest.mark.retry(3)

D. @pytest.mark.flaky(reruns=3)

Solution

1. Step 1: Recall the correct decorator name and parameter

   The correct decorator is @pytest.mark.flaky with parameter reruns to specify retry count.

2. Step 2: Match syntax with options

   @pytest.mark.flaky(reruns=3) uses @pytest.mark.flaky(reruns=3), which is the correct syntax for retrying 3 times.

3. Final Answer:

   @pytest.mark.flaky(reruns=3) -> Option D

4. Quick Check:

   Decorator name + reruns param = correct syntax [OK]

Hint: Use @pytest.mark.flaky with reruns param [OK]

Common Mistakes:

• Using @pytest.retry instead of @pytest.mark.flaky
• Using wrong parameter names like tries
• Missing the mark prefix

3. Given this pytest test code snippet, what will be the output if the test randomly fails once and passes on retry?

import pytest

@pytest.mark.flaky(reruns=2)
def test_random():
    import random
    assert random.choice([True, False])

medium

A. Test may pass after 1 or 2 retries if random returns True

B. Test always fails because random.choice is unpredictable

C. Test runs only once and fails if random returns False

D. Test is skipped due to flaky mark

Solution

1. Step 1: Understand the flaky decorator effect

   The test will rerun up to 2 times if it fails, allowing multiple chances to pass.

2. Step 2: Analyze the random assertion

   The assertion randomly passes or fails. If it fails once, retry can pass if random returns True on retry.

3. Final Answer:

   Test may pass after 1 or 2 retries if random returns True -> Option A

4. Quick Check:

   Retries allow passing despite random failure [OK]

Hint: Retries give multiple chances to pass random failures [OK]

Common Mistakes:

• Assuming flaky skips tests
• Thinking test always fails on first try
• Ignoring retry count effect

4. You wrote this flaky test with retry but it never retries on failure. What is the error?

import pytest

@pytest.mark.flaky(rerun=3)
def test_example():
    assert False

medium

A. The test must return True to retry

B. The decorator should be @pytest.retry, not @pytest.mark.flaky

C. The parameter name should be 'reruns', not 'rerun'

D. The test cannot retry if it always fails

Solution

1. Step 1: Check the decorator parameter spelling

   The correct parameter for retry count is reruns, not rerun.

2. Step 2: Understand impact of wrong parameter

   Using rerun is ignored by pytest, so no retries happen despite failures.

3. Final Answer:

   The parameter name should be 'reruns', not 'rerun' -> Option C

4. Quick Check:

   Correct param spelling = reruns [OK]

Hint: Check exact parameter spelling: reruns, not rerun [OK]

Common Mistakes:

• Misspelling reruns parameter
• Using wrong decorator name
• Expecting retries on always failing test

5. You want to reduce false failures from a flaky test that sometimes fails due to timing issues. Which approach best combines flaky test detection and retry in pytest?

hard

A. Use @pytest.mark.flaky(reruns=3) and add explicit wait in test code

B. Use @pytest.mark.skip to ignore the flaky test

C. Remove retries and fix test to always pass

D. Use @pytest.mark.flaky(reruns=0) to disable retries

Solution

1. Step 1: Understand flaky test retry purpose

   Retries help reduce false failures by rerunning tests that fail randomly.

2. Step 2: Combine retry with test stabilization

   Adding explicit waits addresses timing issues, improving test stability alongside retries.

3. Step 3: Evaluate other options

   Skipping ignores tests, removing retries loses retry benefit, and reruns=0 disables retries.

4. Final Answer:

   Use @pytest.mark.flaky(reruns=3) and add explicit wait in test code -> Option A

5. Quick Check:

   Retries + waits = best flaky test handling [OK]

Hint: Combine retries with waits to fix flaky timing issues [OK]

Common Mistakes:

• Skipping flaky tests instead of fixing
• Disabling retries by setting reruns=0
• Ignoring timing issues causing flakiness