Tagged
On 9 Sep 2026 AWS raised LMI async and ESM timeout from 15 to 90 minutes (5,400 s). A 90-minute SQS consumer needs at least 9 hours of visibility. Sync and classic on-demand Lambda stay at 15 minutes.
On a composite ingest workload (~8k ordered TPS, 1 KB payloads), staying on SQS FIFO without high-throughput mode capped effective throughput near 300 TPS/API and modeled queue backlog cost near $95/mo before ops time — enabling high-throughput FIFO or switching to Kinesis on-demand changed the ceiling, not the consumer code.
EventBridge looks like a $1/million-events service. It is actually six different billing dimensions — custom events, Pipes at $0.40/M, API Destinations at $0.20/M, Schema Discovery at $0.10/M, Archive at $0.10/GB-month, and cross-region replication that doubles the publish line. Built-in AWS-service events are free; custom buses are where the bill lives.
Standard SQS queues sustain nearly unlimited throughput per queue (AWS-documented pattern) while FIFO caps at 300 TPS per API batch without high-throughput mode—your May 2026 architecture review should start from those numbers, not from Kafka slogans.
A technical guide to hybrid compute architectures that combine EC2, Lambda, Fargate, and Step Functions — with worked cost calculations, SQS buffering patterns, and decision frameworks based on invocation pattern rather than unit cost.
SQS, MSK Kafka, and Redis queues are not interchangeable. Each has different cost models, ordering guarantees, and failure modes. This guide covers when to use each, how to autoscale workers on queue depth, and how to build idempotent consumers.
Event-driven sounds elegant in design docs. Then a malformed event ships to a critical bus and you spend a weekend debugging fan-out. EventBridge buses, rules, schema discovery, cross-account patterns, and the failure modes you only learn about in production.