(1/3) One of the hardest problems in the world of stream processing is upgrading stateful applications. Why? Because the events never stop, and your application has to handle the events correctly before and after a upgrade and if rewind your position in a topic.

Should you use Kafka? If so, when? And what are the tradeoffs presented by the dizzying variety of Kafka-adjacent technologies? I hope that my latest blog post provides unique and useful answers to these questions.

Let me know what you all think!

👉 www.responsive.dev/blog/why-whe...

This is just a tiny sample of what's possible when you embrace the LOG as a core building block. Exciting times ahead!

It was so much fun to observe the @responsive.dev team brainstorm distributed state snapshots for Kafka Streams using SlateDB/RS3.

Nothing beats the thrill of having a team of the best engineers solving hard problems with nothing but a whiteboard in front of them!

It's remarkable how popular Kafka Streams is. Here's data for people searching for solutions to their Kafka Streams problems and landing on @responsive.dev. That's an up-and-to-the-right chart I love!

On that note, what's the best book or blog you've read for optimizing website conversions?

I've always wanted to have a landing page for apurvamehta.com. But a blog was too much of a commitment and a short bio was too simple.

@bsky.app allows you to display your feed on a static webpage, which is a perfect setup and is perfectly aligned with the 'your domain is your id' philosophy too!

From the earliest days of @responsive.dev, we knew we'd need a custom database for stream processing. Streaming workloads have too much volume and variability, making existing options impractical at scale.

Last November @rohanpd.bsky.social wrote this while researching our options (2/n)

How about:

It looks like X suppresses posts that link to Bluesky. All my other posts on X get at least 1000 views, except when I link here.

I just enabled ‘threaded mode’ in the web app settings. It’s actually a much nicer way to read a conversation and IMO materially better that X. Try it out!

Operational pain isn't only the reason to drop embedded RocksDB from your stream processor: RocksDB was built for disks, which are expensive in the cloud!

👇🏽 Check out the costs of embedded RocksDB vs traditional disk-centric databases vs a key-value store built with the disk-less SlateDB.