S3 is object storage, vector DB and now File System. 🚀☁️ #100

The AWS Frontier Agents go GA! And we are 100 issues on the newsletter

Apr 15, 2026

In the previous issue, Did we forget Step Functions over Durable Functions?🚀☁️ #99 on how Step Functions is no less in the game of efficient workflows.

In this issue, we are looking into S3 Files and paired with AWS Lambda, ECS (Fargate), and EKS. It effectively turns S3 into a low-latency, “warm” file system for serverless compute without the need for manual data staging.

For serverless, S3 Files turns your "cold" data lake into a "hot" local drive, eliminating the code complexity of managing local caches and data synchronization.

“/tmp pattern” in AWS Lambda involves a three-step dance: downloading an object from S3 to the local ephemeral storage, processing it, and uploading it back. While functional, this introduces latency (cold starts), adds code complexity, and is limited by the 10 GB /tmp ceiling. Amazon S3 Files fundamentally changes this by replacing the Data Transfer phase with a Mount phase and using the file system to easily access files.

⭐ Pick of the month

This month’s Pick is “SQS + Lambda”.

Rehan van der Merwe breaks down the 7 SQS Lambda EMS Scaling behaviors from Lambda invocations and how cold start depends on batch sizes and ESM poller invoking the Lambda.

Luiz Machado explains about the dashboard he has built - SQS Admin Panel: a serverless panel to manage queues, and how to run it locally as well as deploy it with SAM CLI.

Jones Zachariah Noel’s (author) has published how to get started with SQS and understanding types of SQS: Standard and FIFO. If you have cross-account messages, the SQS queue can trigger cross-account Lambda. Also, if you are internested with SQS monitoring and the metrics to look at to understand SQS performance.

🚀 AWSome content to learn from

Aditya Jha breaks down the latest updates with AWS Lambda and must know for developers in 2026 with durable functions, managed instances and runtime flexibility along with lower latency for SQS triggered workloads.

Luis Carlos Osorio Jayk’s CRUD MVPs Starter Kit on AWS Lambda Web Adapter (with Claude Code Support) leveraging Server Side Rendering and Lambda function with Lambda Web Adapter.

Darryl Ruggles’s blog on Amazon Aurora DSQL: A guide to AWS’ Distributed SQL DB that rundowns on the architecture and code with the best practices to go with a truly multi-region DB.

Harpreet Siddhu digs into the 5 Replay Pitfalls in AWS Lambda Durable Functions & How to Design Around Them be it branching decision or static names how it will all play out together!

Eric Johnson shares how Lambda just got a file system and adding AI agents on it brings the best out of S3 Files and adding Strands agent enables the file tools that operates on the file system.

Aashana Shanu explains how using AWS Lambda as an MCP server and the lessons learnt with building AI agents that recover the abandoned carts with the challenges and trade-offs.

Kento IKEDA explains how Amazon S3 Files eliminates the architectural layers and also creates new ones and the real problem S3 is addressing with S3 Files and how it’s more than the mount file system.

Manoj Fernando gets into the event-driven Lambdas with email triggers using Amazon SES explaining how the inbound emails and MX records could be used for triggering Lambda functions.

Luc van Donkersgoed reveals how AWS News leverages semantic content search with the embeddings and S3 vector buckets adding in a good value using Agentic AI and the whole process of how AWS News uses it.