ecs

On a composite B2B order API (~2.4k RPS peak, 24 ECS tasks on Aurora), running the scaling ladder in order — RDS Proxy, read-replica routing, ElastiCache, then SQS offload — moved modeled p99 from 1,840 ms to 72 ms before any microservice split.

AWS shipped ECS Express Mode on November 21, 2025 — three inputs (image + two IAM roles) and Express Mode provisions Fargate, ALB, HTTPS, auto scaling, and a *.ecs.*.on.aws URL. Up to 25 services can share one ALB. No Express Mode surcharge.

ECS CodeDeploy and Lambda aliases support both instant cutover and gradual shifts—but picking wrong costs you double Fargate spend or 21-day MTTR on muted alarms. This decision guide scores blue/green, canary, and rolling with a matrix and names App Mesh (EOL Sept 30, 2026) replacements.

Most DevOps guides teach what AWS services are. Production teaches what happens when 200 engineers use them together. Here's the gap.

ECS blue/green without downtime: CodeDeploy canary/linear still works — and native ECS BLUE_GREEN (2025+) is now a first-class path. July 2026 checklist: hooks, alarms, dual-run cost.

Monolith → Fargate without a big-bang cutover: containerize, ALB weights 10→50→100, keep blue 2–4 weeks. July 2026 cutover checklist.

Autoscaling was supposed to make costs predictable by matching capacity to demand. Instead, it introduced feedback loops, burst amplification, and — with AI workloads — a new class of non-deterministic spend that no scaling policy anticipates.

Autoscaling surprise bills are pattern-shaped: asymmetric thresholds, bad metrics, Lambda duration, Spot storms. July 2026 refresh — target tracking, Budget Actions, FinOps Agent.

A B2B SaaS stack that costs $500/month at launch does not need to cost $50,000/month at 100,000 users if the architecture decisions at each stage are deliberate. This is the end-to-end reference architecture with real cost numbers.

Pick runtimes by ECS density and team skills — not synthetic RPS. July 2026: FrankenPHP worker vs PHP-FPM, Node, Go memory bands + density matrix.

A deep technical guide to running PHP, Python, and Node.js applications on Amazon ECS in production — covering Fargate vs EC2, FrankenPHP vs Nginx+FPM, multi-container task patterns, zero-downtime deployments, and observability.

ECS concurrency: PHP-FPM vs Octane, Node event-loop lag, Gunicorn/asyncio, Go goroutines. July 2026 matrix + ulimit checklist.