Josep Prat

Following our announcement a couple of weeks a go with Inkless Clusters being on Aiven Cloud. We're happy to share that Inkless is now live on GCP. This means that users can now create new Inkless services on GCP, without requiring to be on BYOC. Learn more here: aiven.io/inkless?utm_so…

KIP-1150: Diskless Topics has been officially approved for Apache Kafka. It enables storage and compute separation by writing messages directly to object storage (S3, GCS, Azure Blob), with broker disks acting as optional caches. A big step toward a more cloud-native Kafka architecture. Read more: aiven.io/blog/kip-1150-… #ApacheKafka #StreamingData #CloudNative #DistributedSystems #DataInfrastructure

2025 gave us plenty to celebrate: we passed $100M ARR, launched Inkless Cloud on BYOC, introduced the PostgreSQL Developer Tier, and rolled out Free Kafka - with more to come in 2026.

@kellabyte 😃

Merry Christmas and Happy Holidays to all. I’ve enjoyed sharing, learning, debating and laughing this past year as I returned back to Twitter in 2025. It’s been a blast. 2026 feels like it might be a wild one. Let’s get through it together.

Benchmarking Diskless Topics after launching Inkless on BYOC: 144 producers + 144 consumers, 1 GiB/s in + 3 GiB/s out across 3 AZs. Spoiler: ~1s latency for ~94% lower cost. Read the full report here: aiven.io/blog/benchmark…

🚨 So after 21 years of career, it finally happen, I have been let go from my company, in fact the complete marketing team has been let go ... It came really as a complete surprise, I was not expecting that at all since they were super happy with and the significant impact I had. Also the company is doing really well ... But it's life, it's not easy, I'm recovering slowly and I'm already looking for my next adventure. So if you are looking for a experienced IT professional with expertise going from advocacy (community, pre-sales & customers) to Java, with a lot of knowledge and opinions on Platform Engineering/k8s ecosystem and how all of this can be infused with AI, don't hesitate to ping me. Remote EMEA Also don't hesitate to RT this post. Love you all, my network of friends has already been beyond awesome helping when I told them in private last week.

Listen to The Ravit Show on Spotify for Creators open.spotify.com/episode/0W7CHt… @jlprat @VervericaData @FlinkForward @aiven_io

Aiven Inkless Cloud BYOC is GA: a multi-modal cloud streaming service for Apache Kafka, helping teams running clusters ≲1 MB/s cut complexity and cost. It unifies KRaft, Diskless Topics, and Tiered Storage into a single elastic, open-source API. #Kafka aiven.io/blog/apache-ka…

Speaking in #Barcelona is always special for me! And tomorrow, Wednesday 15th I'll be presenting at Flink Forward Barcelona. Join me and Igor Kersic to hear about The Future of Streaming with @VervericaData and @aiven_io. Come by and say hello.

🚨Kafka Streams 4.1.0 contains a memory leak regression bug: issues.apache.org/jira/browse/KA… Be careful upgrading, or wait for 4.1.1 bug-fix release. We plan to get it out quickly.

I am once again begging you to put your database servers and application servers in the same region.

The last few months of @aiven_io and #ApacheKafka 🧵: - KIP 1150: Diskless Kafka: aiven.io/blog/guide-dis… - Iceberg topics: aiven.io/blog/iceberg-t… - 30 seconds to stream: youtube.com/watch?v=wGqOX9…

Congratulations to @FrankYangTW for becoming a committer and to @lucasbrut for being added to the PMC!

The largest Kafka you've never heard about: Datadog 🐶 • hundreds of TRILLIONS of messages a day • terabytes a second of traffic - at least 10x more than Uber's 🔥 • >10,000 brokers in 600 Kafka clusters • >1,000,000 partitions over thousands of topics - e.g. 7000 topics with ~140 partitions each At this scale, they don't just use Kafka - they've built a whole Streaming Platform around it. You can call it Kafka 2.0, because it might as well be its own product. It includes things like: 🦀 Custom Clients - their own Kafka client built in Rust, alongside additional business logic like live traffic failover (redirecting traffic to a different cluster) 🧭 Custom Consumer Group Coordinator - one that works across clusters. It also monitors and triggers rebalances out of band in seconds (not using the slow Kafka Consumer Group protocol) 🧠 Custom Consumer Commit Log - allows it to track multiple offsets per partition simultaneously and includes real-time wall-clock consumer lag tracking. The system automatically detects growing lag and proactively rebalances it to avoid further accumulation (while processing the leftover in the background) 🌀 Proactive Rebalancing - the Streaming Platform also does its own proactive rebalancing - distributing traffic, changing partition counts and decommissioning clusters freely 🎪 Cells - they have separated brokers into cells (called Streams) spanning different Kafka clusters 🚦 QoS lanes - they split topics into lanes with different QoS (quality of service) guarantees - regular lane, slow lane and a DLQ lane An interesting trend I see here as well as in other digital-native businesses (e.g OpenAI & Stripe) is opting out of strict ordering and strict delivery guarantees in the interest of simplicity and scalability ✨ Datadog uses an at-least-once delivery model and has relaxed ordering guarantees in Kafka - the processing stage of the pipeline. Instead, they re-establish order later in a downstream system. 👌 Kudos to an absolute powerhouse of a Kafka team 👏

A Conversation with the Experts: Real-Time, Real Easy Streaming with DigitalOcean Managed Kafka x.com/i/broadcasts/1…

Questions about Kafka? We've got the answers. 💬☁️ Join Kafka experts @jlprat & @FvYonov on May 28 as they discuss real-world use cases, practical architecture tips, & answer questions from the audience. 🔗 do.co/3ZrhSbR

I spoke on another podcast about KIP-1150 (diskless topics in Kafka). Thanks @kozlovski and @gharris1727 !

BLOGGED 🎉 - Diskless Kafka KIP-1150 Wrote a detailed blog on why I think diskless Kafka is going to be HUGE. Do check it out here: geeknarrator.com/blog/diskless-… @aiven_io has been at the forefront of this proposal that is going to help the entire #Kafka ecosystem. Also, thanks @aiven_io for sponsoring the blog.

I just published a 2 hour and 30 minutes long technical deep dive podcast with two of the Diskless proposal authors. The podcast is a treasure trove chock full of deep technical insights. There is no better resource online about KIP-1150 🔥 👉 youtube.com/watch?v=hrMvOF…