---
title: "Comment by Strazi Weekey on \"A Deep Dive into AWS Load Balancers: CLB, ALB, NLB, and GWLB\" | Hashnode"
description: "Great deep dive. I recently migrated a legacy app from a CLB to an NLB for its TCP-preserving features, and the performance gain for our WebSocket connections w"
url: "https://hashnode.com/posts/a-deep-dive-into-aws-load-balancers-clb-alb-nlb-and-gwlb/67e7d695d57bae15082d8ae5/comment/69d661e4a74bb14986a54b8d"
---

[Discussion](https://hashnode.com/posts/a-deep-dive-into-aws-load-balancers-clb-alb-nlb-and-gwlb/67e7d695d57bae15082d8ae5)

Comment [A Deep Dive into AWS Load Balancers: CLB, ALB, NLB, and GWLB](https://hashnode.com/posts/a-deep-dive-into-aws-load-balancers-clb-alb-nlb-and-gwlb/67e7d695d57bae15082d8ae5)

[Strazi Weekey](https://hashnode.com/@weekey)

[Apr 8](https://hashnode.com/posts/a-deep-dive-into-aws-load-balancers-clb-alb-nlb-and-gwlb/67e7d695d57bae15082d8ae5/comment/69d661e4a74bb14986a54b8d)

Great deep dive. I recently migrated a legacy app from a CLB to an NLB for its TCP-preserving features, and the performance gain for our WebSocket connections was exactly as you described—night and day. Your breakdown of the "amateur hour" single-server trap is spot on.