behave-retry 1.0.0

Automatic retry for failed Behave scenarios — CLI flags, tag overrides, exception filtering, flakiness stats

behave-retry

Automatic retry for failed Behave scenarios — CLI flags, tag overrides, exception filtering, and flakiness stats.

Why?

Behave has no built-in retry. When a scenario fails due to flakiness (timing, network, race conditions), there's no way to re-run it automatically. Cucumber has --retry natively. Behave doesn't.

behave-retry fills that gap with a lightweight, zero-dependency library that integrates through Behave's hook system.

Table of contents

• Install
• Quick start
• Features
  • Global retry
  • Tag-filtered retry
  • Exception-filtered retry
  • Per-scenario override
  • Cleanup between retries
  • Retry stats
• API reference
• Configuration
• Examples
• Contributing
• License

Install

pip install behave-retry

For development:

pip install -e ".[dev]"

Quick start

# environment.py
from behave_retry import setup_retry, after_scenario_hook, retry_report

def before_all(context):
    setup_retry(context, max_retries=3)

def after_scenario(context, scenario):
    after_scenario_hook(context, scenario)

def after_all(context):
    print(retry_report(context))

That's it. Failed scenarios will now be retried up to 3 times automatically.

Features

Global retry

setup_retry(context, max_retries=3)

Retry every failed scenario up to 3 times.

Tag-filtered retry

setup_retry(context, max_retries=3, retry_tags=["@flaky"])

Only retry scenarios tagged with @flaky.

@flaky
Scenario: Login with slow network
  Given the server is slow
  ...

Exception-filtered retry

setup_retry(
    context,
    max_retries=3,
    retry_on=[AssertionError, TimeoutError],
)

Only retry when the scenario fails with specific exception types. Subclasses of the listed exceptions also match.

Per-scenario override

Override the global retry count per scenario using the @retry:N tag:

@retry:5
Scenario: Very flaky test
  ...

@retry:0
Scenario: Never retry this
  ...

The first @retry:N tag wins. @retry:0 disables retry for that scenario.

Cleanup between retries

Define an after_retry hook in your environment.py to clean up state between retry attempts:

# environment.py
def after_retry(context, scenario):
    # Close browser, reset DB, clean state
    if hasattr(context, "driver"):
        context.driver.quit()

Retry stats

from behave_retry import retry_report

def after_all(context):
    report = retry_report(context)
    print(report)

Output:

Retry Summary:
  Total retries: 5
  Scenarios retried: 3
  Passed on retry: 2
  Failed after retry: 1

  - "Login with invalid credentials" — 3 attempts, failed (AssertionError)
  - "Search products" — 2 attempts, passed
  - "Checkout flow" — 1 attempt, passed

API reference

setup_retry(context, max_retries=0, retry_tags=None, retry_on=None)

Configure retry on the behave context. Call this in before_all.

Parameter	Type	Default	Description
context	Any	—	Behave context object
max_retries	int	0	Maximum retries per scenario (0 = no retry)
retry_tags	list[str] | None	None	Only retry scenarios with these tags
retry_on	list[type[Exception]] | None	None	Only retry on these exception types

after_scenario_hook(context, scenario)

Handle retry logic. Call this in after_scenario.

retry_report(context) -> str

Get a human-readable retry summary. Call this in after_all.

RetryConfig

Dataclass for configuration.

Attribute	Type	Default	Description
max_retries	int	0	Maximum retries per scenario
retry_tags	list[str]	[]	Tag filter (empty = retry all)
retry_on	list[type[Exception]]	[]	Exception filter (empty = retry on any)

Methods:

• should_retry_tag(tags) -> bool — Check if scenario tags allow retry
• should_retry_exception(exc) -> bool — Check if exception type allows retry
• get_scenario_retries(tags) -> int — Get max retries for a scenario, checking @retry:N override

RetryStats

Aggregate retry statistics.

Property	Type	Description
total_retries	int	Total retry attempts across all scenarios
scenarios_retried	list[ScenarioRetry]	Per-scenario retry records
scenarios_passed_on_retry	int	Scenarios that passed after retry
scenarios_failed_after_retry	int	Scenarios that still failed after retry

ScenarioRetry

Record of retry attempts for a single scenario.

Attribute	Type	Description
scenario	str	Scenario name
attempts	int	Total attempts (including first run)
final_status	str	"passed" or "failed"
exceptions	list[str]	Exception types encountered

Property	Description
was_retried	True if retried at least once
passed_on_retry	True if passed after retry

RetryExhaustedError

Raised when a scenario has been retried the maximum number of times.

parse_retry_tag(tags) -> int | None

Parse @retry:N tag from scenario tags. Returns N if found, None otherwise.

Configuration

Python version

behave-retry requires Python 3.11+.

Dependencies

Zero required dependencies. behave is only needed as a dev dependency for running tests.

Linting and formatting

The project uses Ruff with the following rule sets: E, F, W, I, N, UP, B, SIM.

ruff check .

Testing

pytest tests/ -v --cov

Coverage threshold is 90%.

Examples

Full environment.py

from behave_retry import setup_retry, after_scenario_hook, retry_report

def before_all(context):
    setup_retry(
        context,
        max_retries=3,
        retry_tags=["@flaky"],
        retry_on=[AssertionError, TimeoutError],
    )

def after_scenario(context, scenario):
    after_scenario_hook(context, scenario)

def after_all(context):
    print(retry_report(context))

Feature file with retry tags

@flaky
@retry:5
Feature: Flaky scenarios

  @retry:0
  Scenario: Never retry this
    Given a stable condition
    Then it should pass

  Scenario: Retry up to 5 times
    Given a flaky condition
    Then it might fail

Contributing

Contributions are welcome! See CONTRIBUTING.md for guidelines.

License

MIT — Copyright (c) 2026 Mathias Paulenko