Apache Kafka®️ 비용 절감 방법 및 최적의 비용 설계 안내 웨비나 | 자세히 알아보려면 지금 등록하세요

What’s New in Confluent Clients for Kafka: Python Async GA, Schema Registry Upgrades

작성자:

Hey, fellow Apache Kafka® developers! It’s time for another update on the Confluent client ecosystem. Following our recent architectural milestones, we’re excited to announce the release of librdkafka 2.13.0, which powers the latest versions of our Python, JavaScript, .NET, Go, and C/C++ clients.

In this release, you’ll find numerous improvements to the Python experience as well as critical security and Schema Registry enhancements for everyone.

Python Developers: asyncio Is Now GA

The headline for the Python community is that asyncio support has officially reached General Availability (GA). After a successful preview period, we’ve refined the interface to make it the gold standard for modern, asynchronous Python applications.

What’s improved in Python?

  • Better Life Cycle Management: Enhanced context managers (with statements) and cleaner shutdown behavior using close() make resource management more intuitive.

  • Richer Metadata: We’ve introduced richer Message objects and deterministic partitioners to give you more control over your data flow.

  • Developer Experience (DevEx): We closed more than 100 GitHub issues in the Python repository. This includes adding comprehensive type hints and tightening up the experience for long-running workloads.

from confluent_kafka import Consumer
consumer_conf = {
	'bootstrap.servers': 'localhost:9092',
	'group.id': 'my-group',
}

with Consumer(consumer_conf) as consumer:
	consumer.subscribe('my-topic')
	msg = consumer.poll(timeout=1.0)
print(f"value={msg.value().decode('utf-8')}")

Whether you’re building with FastAPI or custom async logic, the Python client is now more robust and "Pythonic" than ever.

Enhancing the Schema Registry Experience

For those relying on Stream Governance, we’ve implemented a suite of cross-language upgrades to make using Schema Registry smoother and more predictable. These updates apply across Python, .NET, Go, and JavaScript.

  • Apache Avro Improvements: We now have better support for Avro schema references and associations.

  • Stricter Validation: New validation flags help catch schema mismatches earlier in the development cycle.

  • Bug Fixes: We’ve addressed several edge cases involving wrapped unions, bytes serialization, and caching to ensure that your data remains consistent across different languages.

Security, Stability, and Performance

The core engine, librdkafka, received several under-the-hood improvements that benefit the entire ecosystem simultaneously.

  • Future-Proofing Admin APIs: We’ve updated support for KIP-482, upgrading key admin APIs to improve compatibility with newer Kafka brokers and to ensure long-term stability for administrative operations.

  • Safety First: This release includes critical fixes for SSL handling, thread safety, and memory safety.

  • Performance Tuning: We’ve implemented targeted performance improvements specifically for .NET, Go, and JavaScript to ensure that high-throughput workloads remain stable under pressure.

  • Reliable Auth: Improved OAuth token refresh handling ensures that your secure connections aren't interrupted by credential expiration issues.

Summary of Key Updates

Here’s a quick snapshot of what’s landing in the 2.13.0 release:

  • Python: GA for asyncio, with full type hints, better shutdown handling, and 100+ GitHub issues closed

  • Schema Registry: Improved Avro reference support and stricter validation across all non-Java clients

  • Admin APIs: Enhanced KIP-482 support for better compatibility

  • Ecosystem Stability: Performance and safety fixes for .NET, Go, and JavaScript

What’s Next for Confluent’s Kafka Clients

We’re not slowing down. While this release leaned heavily into the Python ecosystem, our next phase involves broad improvements across the board, with planned deep dives into librdkafka core, .NET, and JavaScript clients.

As always, we thrive on your input. Please continue to funnel your feedback and feature requests into our community channels. Happy streaming!


Apache®, Apache Kafka®, Kafka®, Apache Avro, and Avro are registered trademarks of the Apache Software Foundation in the United States and/or other countries. No endorsement by the Apache Software Foundation is implied by using these marks. All other trademarks are the property of their respective owners.

  • Jaisen Mathai is a Product Manager at Confluent, leading the strategy for the official Apache Kafka clients. With over 25 years of industry experience, Jaisen brings a unique perspective, having spent time as both an engineer and a product leader at companies including Google. Jaisen's focus is ensuring high-quality client experiences for developers building with Kafka in C/C++, .NET, Javascript, Go, and Python.

이 블로그 게시물이 마음에 드셨나요? 지금 공유해 주세요.