Bulkman

Enterprise-Grade Bulkhead Pattern for Python

Bulkman is a robust, production-ready implementation of the Bulkhead pattern, designed to isolate resources and prevent cascading failures in mission-critical distributed systems. It supports both modern async (AnyIO/asyncio) and traditional synchronous (Threading) workloads with zero overhead.

Key Capabilities

• Resource Isolation: Limits concurrent access to prevent service exhaustion.
• Circuit Breaker Integration: Built-in distributed circuit breaking via resilient-circuit (opt-in, disabled by default).
• Dual Mode:
  • Async: AnyIO-based (asyncio) for high-concurrency IO.
  • Sync: Zero-overhead ThreadPoolExecutor implementation for CPU-bound or blocking IO tasks.
• Observability: Full contextvars support for distributed tracing (OpenTelemetry/Datadog).
• Type Safe: 100% type-hinted and rigorously tested (>92% coverage).

Documentation

• User Guide: Installation, Quick Start (Async), and Configuration.
• Synchronous / Threading Guide: Best practices for blocking workloads (Flask, Django, Scripts).
• Contributing: Setup, testing, and contribution guidelines.

Quick Install

pip install bulkman

Quick Example (Blocking/Threading)

For legacy or synchronous applications, use BulkheadThreading:

from bulkman import BulkheadThreading, BulkheadConfig

# Configure
config = BulkheadConfig(name="db_writer", max_concurrent_calls=10, timeout_seconds=5.0)
bulkhead = BulkheadThreading(config)

# Execute
try:
    future = bulkhead.execute(my_blocking_function, data)
    result = future.result(timeout=5.0)
    print(result.result)
except TimeoutError:
    print("System overloaded")

👉 See the User Guide for async examples

License

Licensed under the Apache License 2.0.