Best of AWS — April 2026

  1. 1
    Article
    Avatar of bytebytegoByteByteGo·4w

    EP210: Monolithic vs Microservices vs Serverless

    A curated system design newsletter covering five topics: a comparison of monolithic, microservices, and serverless architectures with practical tradeoffs; CLI vs MCP for AI agents across six dimensions including token cost, auth, and governance; a comparison of five major coding agents in 2026; an overview of essential AWS services and where they fit in production stacks; and a beginner-friendly visual explanation of JWT structure and security.

  2. 2
    Article
    Avatar of sdtimesSD Times·3w

    Introducing enhancements in Kiro CLI 2.0

    AWS Kiro CLI 2.0 has been released with three major additions: headless mode for running in CI/CD pipelines and automated workflows without a browser (via API key), native Windows support for use in Windows Terminal, and a refreshed terminal UI featuring a subagent experience with progress monitoring, an updated todo list, and a new task list.

  3. 3
    Article
    Avatar of developingdevThe Developing Dev·4w

    AWS Distinguished Eng: Learnings From 3000 Incidents And How Engineering Is Changing

    Marc Brooker, AWS Distinguished Engineer, shares insights from reading 3,000+ cloud system postmortems, covering what makes great postmortems, why on-call is a powerful learning tool, and how AWS's weekly COE review has been central to its success. He explains why caches can be dangerous in distributed systems due to metastable failures, and how Aurora DSQL was designed to avoid common relational database outage patterns using MVCC and optimistic locking. He also shares his perspective on how AI will reshape software engineering careers, advising junior engineers to focus on understanding customers and problems, and senior engineers to stay hands-on with modern agentic tools. He also advocates for writing as a tool for both scaling expertise and sharpening thinking.

  4. 4
    Article
    Avatar of mondaymonday Engineering·3w

    Building a Reliable and Extendable Notifications Platform

    monday.com's engineering team shares how they rebuilt their notifications platform from a Ruby on Rails monolith into a scalable TypeScript microservice. The new system uses a three-stage SQS-based orchestration pipeline (processing, filtering, delivery) with exponential backoff, per-channel retry logic, and DLQ support. A JSON-based app feature format and SDK allow developers to add new notification types in about an hour. The platform now handles over 8 million notifications daily across email, mobile, Slack, MS Teams, and in-app channels, with Datadog instrumentation for observability.

  5. 5
    Article
    Avatar of collectionsCollections·4w

    How S3 Files adds real file system access to S3 buckets

    AWS launched S3 Files, a feature that mounts any S3 bucket or prefix as a shared NFS v4.1/4.2 filesystem on EC2, ECS, EKS, Lambda, and Fargate. Unlike previous workarounds (s3fs-fuse, goofys, Mountpoint for S3), it is built on Amazon EFS infrastructure with S3 as the durable backing store. A two-tier caching model keeps hot data in EFS at ~1ms latency while large sequential reads stream directly from S3 at up to 3 GB/s. A 'stage and commit' design enables full POSIX semantics including in-place edits, file locking, and directory renames. Pricing mirrors EFS Performance-optimized mode at $0.30/GB-month for hot storage, with large-file reads bypassing S3 Files charges entirely. Key use cases include ML training pipelines, agentic AI workloads needing concurrent access, and legacy POSIX apps on EFS seeking S3 storage economics. Available now in all commercial AWS regions.

  6. 6
    Article
    Avatar of phoronixPhoronix·5w

    AWS Engineer Reports PostgreSQL Performance Halved By Linux 7.0, But A Fix May Not Be Easy

    An AWS engineer discovered that the Linux 7.0 development kernel cuts PostgreSQL throughput roughly in half on Graviton4 hardware. The regression was traced to Linux 7.0 restricting available kernel preemption modes, causing significantly more time spent in user-space spinlocks. A patch was proposed to restore PREEMPT_NONE as the default, but kernel maintainer Peter Zijlstra pushed back, arguing the proper fix is for PostgreSQL to adopt the Restartable Sequences (RSEQ) time slice extension introduced in Linux 7.0. If the patch is rejected, PostgreSQL users could face major performance degradation when Linux 7.0 stable ships in approximately two weeks.

  7. 7
    Article
    Avatar of awsfundamentalsAWS Fundamentals·3w

    AWS S3 Files vs EFS — Cost Comparison, POSIX Limits, and Lambda Setup

    AWS S3 Files is a managed POSIX/NFS file system backed by S3 storage, offering up to 13x lower storage costs than EFS ($0.023/GB vs $0.30/GB) for large-file workloads. The post covers how S3 Files works (file system, mount target, access point), the full cost comparison including edge cases where S3 Files is actually more expensive (small files under 1 MiB), the critical 60-second write-back delay that breaks coordination patterns, POSIX limitations (no hard links, no cross-client file locking, non-atomic directory renames), and a step-by-step AWS SAM setup for Lambda including VPC networking, service role, and file system provisioning. A decision table helps choose between S3 Files and EFS based on file size, consistency needs, and concurrency requirements.

  8. 8
    Article
    Avatar of giantswarmGiant Swarm·5w

    Live migrating hundreds of Kubernetes clusters to Cluster API

    Giant Swarm replaced their custom-built Kubernetes cluster management system with Cluster API (CAPA), live-migrating hundreds of enterprise AWS production clusters without downtime or data loss. The post details the technical mechanics: a CLI-based migration tool, a two-phase process covering CR migration and node transition, and a creative workaround involving forking HashiCorp Vault to extract root CA signing keys for certificate continuity. Key lessons include the importance of aligning team structure with stated priorities, avoiding premature expansion to new providers before completing the core migration, and the strategic value of adopting upstream open source when a custom solution is no longer differentiating. The migration took years, required company-wide involvement, and ultimately freed engineering capacity for higher-value work.

See all AWS archives