jetstream VS kafkacat

Centrifugo can use Nats as PUB/SUB broker. I also tried to use Jetstream to implement Centrifugo history/recovery semantics – unfortunately Jetstream does not fit us very well as streams must be created over a separate call to admin API, and it's not possible to iterate/paginate over Jetstream streams to be suitable to current Centrifugo message history and recovery semantics (I asked about support for this – unfortunately the issue was closed and required API not implemented - https://github.com/nats-io/jetstream/issues/266, suggested solution is far from optimal for our use case).

To communicate with Kafka, you can use Kafkacat, a command-line tool that allows to produce and consume Kafka messages using a very simple syntax. It also allows you to view the topics' metadata.

Feels like you could get pretty far with kafkacat and a SQLite database.

This is the second release version of Kafcat. Kafcat is a Rust fully async rewrite of kafkacat.

Consume from the Kafka topic my.topic with kafkacat:

From the code above, you can see that to process the headers, simply use the ConsumerRecord.headers() method to return the headers. In our example above, we’re printing the headers out to the console for demonstration purposes. Once you have access to the headers, you can process them as needed. For reading headers from the command line, KIP-431 adds support for optionally printing headers from the ConsoleConsumer, which will be available in the Apache Kafka 2.7.0 release.You can also use kafkacat to view headers from the command line. Here’s an example command:

Note: In the example above, we have used kafkacat to consume the topics. The option -o-1 is used to only consume the latest message