confluent-kafka-python VS liftbridge

Responding to popular demand, the latest 0.4.0rc0 version introduces support for Kafka stream processing using Confluent Kafka's Python library - https://github.com/confluentinc/confluent-kafka-python.

- confluent kafka library (https://github.com/confluentinc/confluent-kafka-python) uses librdkafka, and is faster that kafka-python - https://www.bytewax.io/(https://github.com/bytewax/bytewax) is another option. Handles the dirty work of exchanging data across processes for parallel stateful processing

Here are examples https://github.com/confluentinc/confluent-kafka-python

I just wanted to mention that pykafka is currently unmaintained and archived on GitHub:

https://github.com/Parsely/pykafka

pykafka was originally developed and maintained by my team at Parse.ly, but we no longer maintain it. We instead encourage folks to use confluent-kafka-python, which is what we have ourselves switched to in our production systems:

https://github.com/confluentinc/confluent-kafka-python

(pykafka was developed at a time before Confluent invested in their own Python binding. Some of the history of the project is described in this 2016 blog post[1] and our original 2015 announcement[2].)

[1]: https://blog.parse.ly/pykafka-now/

[2]: https://blog.parse.ly/announcing-pykafka-python-support-for-...

confluent-kafka

I also found that there is a library provided by confluent https://github.com/confluentinc/confluent-kafka-python which maybe helpful here. Is that right? I see a section "SSL certificates" but couldnt get much insight out of this.

ProtobugDeserializer does not allow anything but the protoBuf message type (see here):

I'd start off with some simple producer/consumer code to get an understanding of interacting with Kafka's APIs from Python - there's a Python client with some good examples of the producer and consumer as well as more complex examples.

Liftbridge (https://liftbridge.io/) is more or less the NATS Kafka versioning.

[Gufo Liftbridge](https://pypi.org/project/gufo-liftbridge/) is the Python asyncio Liftbridge client.

https://liftbridge.io/

Apache Pulsar might be worth a look, but it's actually more complex under the hood than Kafka, but has a lot of features built-in that either aren't in FOSS Kafka yet, like tiered storage, or won't be until Confluent doesn't dominate the PMC (like an integrated schema registry), or just can't be done very nicely, if at all, like decent multi-tenancy.

That said, it's a fast moving target, the code quality last I looked was patchy in places, ditto the documentation for both it and Bookkeeper, and the admin overhead is higher (managing bookies and brokers and Zookeepers vs. just brokers and ZK with Kafka, or when KRaft is production ready, just brokers).

This is basically an ad for GCP right?

That said, it looks like Kafka is by far and away the way to handle persistent logs/events at scale. AFAIK a company here in Japan called LINE has all their messaging flowing through a large kafka cluster themselves.

Wonder if anyone is running large NATS Jetstream[0]/Liftbridge[1] or Pulsar[2] (yahoo runs those) clusters. I guess Pulsar might be #2 in terms of adoption at large scale?

[0]: https://docs.nats.io/jetstream/jetstream

[1]: https://liftbridge.io/

[2]: https://pulsar.apache.org/

Do you need the messages to be durable, if so you can have a look at Liftbridge: - https://liftbridge.io/ - https://github.com/liftbridge-io/liftbridge

And if you want closer kafka semantics built on top of nats, check out liftbridge:

https://liftbridge.io/