Vineel Vallepu

If you’ve worked with Apache Kafka, you know the trade-offs: Blazing-fast local disk writes but costly replication across zones. Tiered storage for long retention but complexity in managing hot vs cold data. The Kafka community has been asking: Can we make object storage the first-class durable store without losing Kafka’s semantics? That’s where Diskless Topics (KIP-1150 & KIP-1163) by Aiven started the journey — rethinking Kafka’s durability around object storage as the source of truth. Now comes Diskless 2.0 — a unified, zero-copy approach that: ✅ Removes parallel “disk vs tiered” estates ✅ Simplifies the produce/consume path ✅ Enables zero-copy migration (flip a topic from classic to diskless without reprocessing) ✅ Cuts costs by eliminating cross-AZ replication and reducing broker disk reliance This isn’t just an incremental patch. It’s a paradigm shift in how we think about Kafka in the cloud era — where durability and economics favor object storage, and Kafka brokers can finally be “lighter.” Of course, challenges remain: -Latency sensitivity (classic still wins for ultra-low-latency cases) -Compaction and deletion semantics (work in progress) -Operational readiness for large clusters But that’s the beauty of open source — progress through community iteration. And Diskless 2.0 feels like a community win, not just a vendor push. 👉 Dive deeper: Read the KIP-1150 proposal here: fandf.co/3VlQxW1 Explore Aiven’s take on Diskless 2.0: fandf.co/46pczw0 What do you think? Could Diskless 2.0 become the default way to run Kafka in the next 3–5 years, or will classic/tiered still dominate latency-critical workloads? Kudos to Aiven for releasing Diskless 2.0 and sponsoring this post. #Kafka #Sponsored #Aiven

266

49 Comments

Here's another part in the series of deploying a multi-agent solution that uses Google Agent2Agent to Azure Container Apps. It's a follow-up to deploying the infrastructure parts. In this post, I deploy the actual containers: ✅ Deployment of the conversational UI (Semantic Kernel) ✅ Deployment of the MCP Server (FastMCP) ✅ Deployment of the Web Agent (OpenAI Agents SDK + A2A) ✅ Deployment of the Rag Agent (private Azure AI Foundry + A2A) Here's the post: https://lnkd.in/e88STKR6 Here's the repo: https://lnkd.in/eEDcwwrK #ai #azure #containerapps #a2a #foundry

104

3 Comments

Postgres’ JSONB attempts to integrate NoSQL flexibility into SQL but encounters severe performance limitations for real-world workloads. MongoDB is the only DB that gives you the flexibility & scale of NoSQL with the reliability of SQL by design, not as a workaround. https://lnkd.in/eXGPEmjx

595

4 Comments

Excited to share our latest work on building Multi-Agent Site Reliability Engineering Agent with Amazon Bedrock AgentCore! In this comprehensive blog post, my co-author Dheeraj Oruganty and I demonstrate how to leverage Amazon Bedrock AgentCore for building an AI Agent that leverage different primitives provided by Amazon Bedrock AgentCore. We use AgentCore Runtime to deploy the multi-agent system (built with LangGraph), AgentCore Gateway to build an MCP server for providing access to the tools that the Agent uses, AgentCore Memory for storing user preferences, semantic facts and conversation summaries, AgentCore Observability for metrics and traces, and of course AgentCore Identity for providing AuthN and AuthZ. The system uses synthetic data to mimic a production system and provides several examples that an SRE would encounter as part of their day to day work. This blog post is meant not only as a showcase of Amazon Bedrock AgentCore's broad capabilities and usefulness but also to provide a real-life example of how AI Agents are solving important business problems. This example was built in response to a request from one of our customers when we were showcasing the capabilities of Amazon Bedrock AgentCore to them. Please reach out if you are interested int customizing this example to work for your systems and see how it can help your SREs! Meanwhile, here are some useful links: 𝗗𝗲𝗺𝗼 𝘃𝗶𝗱𝗲𝗼𝘀: https://lnkd.in/e8kZ-prW 𝗚𝗶𝘁𝗛𝘂𝗯 𝗿𝗲𝗽𝗼: https://lnkd.in/eqtYDraj 𝗕𝗹𝗼𝗴 𝗽𝗼𝘀𝘁: https://lnkd.in/eS4mQZKW #AmazonBedrock #GenerativeAI #SiteReliability #MultiAgentSystems #AWS #MachineLearning #DevOps #AIOps #AgentCore #LangGraph

309

9 Comments

It looks like there are two pieces of news to come out of HashiConf for those of us in the IaC world: Terraform Search (https://lnkd.in/g9PujUDq) -- This essentially gives providers the ability to search against APIs and avoid the "Find the ARN, add to import block, run an apply to finalize import" loop. We do this today at Masterpoint with Claude Code / Cursor Agent mode and that workflow is damn good, but I don't mind seeing this become a first class citizen :thumbsup: Terraform Actions (https://lnkd.in/gpfMwJDP) -- This adds a new action block that allows your TF workload to invoke an action in a downstream tool. Their example is invoking AWS Lambda functions, which you can do via a special apply -invoke=<ACTION> CLI argument OR you can setup triggers so that when a resource changes in your state graph the action will get invoked. Interesting and potentially good functionality from Hashi -- I'm happy to see it. Is it super innovative? I wouldn't say so, but search sounds like a good thing to have as a first class citizen and I like that. It plays in well with the IBM acquisition for sure as one of their main quoted mentions for actions is invoking an Ansible playbook. I personally like that this is part of TF Core, good on them for building it and not limiting it to TFC / TFE. Now let's see how fast it takes OpenTofu to build similar OR come up with their own new additions 🙌 #terraform #opentofu #infrastructure #infrastructureascode #iac #platform #platformengineering #platforms #cloud #devops

47

10 Comments