
My fullstack AWS tech stack in 2026

The AWS tech stack I use for every fullstack project in 2026 - Lambda with Hono, Postgres, TanStack Start, CDK, EventBridge, and SQS - plus what it costs and why I stopped changing it.
Blog
Learn AWS concepts, best practices, and real-world implementations.
Browse by service


The AWS tech stack I use for every fullstack project in 2026 - Lambda with Hono, Postgres, TanStack Start, CDK, EventBridge, and SQS - plus what it costs and why I stopped changing it.


DynamoDB now has native vector search. Learn how to store Titan embeddings next to your items, create a vector index, and use the SearchVectors API to build a related-posts feature without a separate vector database or sync pipeline.


Developers keep shipping s3:* and AdministratorAccess, and it is not laziness. The deploy loop, indirect permissions, policy size limits, and Access Analyzer gaps all push you toward wildcards. Here is what breaks and the account-level setup that works instead.


Stop getting paged at 3am for regional Lambda throttling. Learn why ClaimedAccountConcurrency is the right metric to watch, how to set a CloudWatch alarm at 70% utilization, and how to automate quota increase requests.


Most teams should stay on Lambda and Fargate. At scale, with many teams and a platform org, EKS with Flux, Helm, and Karpenter becomes the right call. Here's the GitOps starter blueprint and the mental model behind it.


A point-by-point response to the viral Hacker News post about leaving AWS. Most of the complaints are skill issues, but a few are fair. Here is my honest take from someone who has run AWS at scale for years.


The AgentCore harness preview pairs a managed runtime, server-side conversation memory, per-session microVMs, multi-provider model support, and streamed observability so you can ship an agent on AWS without writing the orchestration loop.


A walkthrough of the AWS Fundamentals Organization built with Terraform. Service Control Policies, Resource Control Policies, CloudTrail audit pipeline, email and Slack alerts, and a budget-driven kill switch for sandbox accounts.


In this post, you learn the different levels you can have with AWS Observability, from unstructured logs up to a Claude Code skill which finds logs for you automatically.


AWS DevOps Agent is an AI-powered SRE that investigates incidents autonomously. Here's how it works, what it can do, where it falls short, and whether you should try it.


AWS S3 Files brings a managed POSIX file system at a fraction of EFS costs. Learn how it works, where the 60-second write-back delay bites you, and how to set it up with Lambda using AWS SAM.


You handed out AWS and Azure credentials for a 3-week sprint. That was months ago. Here is how to build a self-service access portal that provisions temporary credentials and cleans them up automatically.
Showing 12 of 162