Backon: The Zero-Dependency Retry Library That Leaves Backoff in the Dust

Retry logic in Python is getting a much-needed overhaul. Backon, a new library for Python 3.10+, isn't just nipping at the heels of the venerable backoff, it's leaping over it entirely. Here's why Backon deserves your attention, even if you're perfectly happy with your current retry setup.

The Zero-Dependency Gambit

Backon's boldest move? Ditching dependencies entirely. In an era of supply chain attacks and version conflicts, Backon's stdlib-only approach isn't just neat, it's a strategic advantage. Your retry logic is now one less moving part in your dependency graph.

Four Ways to Retry

Backon offers a Swiss Army knife of APIs:

1. Decorators (@on_exception, @on_predicate)
2. Functional (retry())
3. Context managers (Retrying)
4. Callable objects (RetryingCaller/AsyncRetryingCaller)

This isn't feature bloat; it's recognition that different codebases have different idioms. Backon plays nice with all of them.

Async Without Asterisks

Backon's async support isn't bolted on, it's baked in. The same API handles both sync and async code, eliminating the cognitive overhead of juggling two mental models:

async with backon.Retrying(backon.constant, exception=ValueError, max_tries=3, interval=0.5) as r:
    result = await r.async_call(fetch_data)

Beyond Basic Backoff

Where Backon truly shines is in its advanced features:

• Circuit breakers to prevent cascading failures
• Request hedging for latency-sensitive operations
• Rich callbacks for fine-grained control
• Composable wait strategies for tailored backoff algorithms

These aren't just bells and whistles. They're the tools you need to build truly resilient systems in complex, distributed environments.

The Developer Experience Dividend

Backon isn't just about features; it's about fit. Full type hints, a dedicated testing module, and optional metrics integrations make it feel less like a library and more like a natural extension of your development workflow.

The Upgrade You Didn't Know You Needed

Backon represents a fundamental rethink of what a retry library should be in 2023. It's not just more powerful than backoff, it's more thoughtful. The zero-dependency approach, the unified sync/async API, and the advanced resilience patterns all point to a library designed for the complexities of modern distributed systems.

Yes, the Python 3.10+ requirement might give some teams pause. And yes, there's a learning curve to fully leverage its power. But these are small prices to pay for a retry solution that's truly future-proof.

Backon isn't just a better backoff. It's a challenge to rethink how we approach transient failures in our code. For teams pushing the boundaries of what's possible with Python, Backon isn't just an option, it's quickly becoming the obvious choice.