I'm trying to find a way to make Target Tracking scaling for CPU on ECS Fargate more responsive. Currently, the CloudWatch alarms are set to trigger scale-outs based on three metrics over a period of 60 seconds, but the scale-in period is much longer, which is causing downtime for my application when CPU usage spikes. Is there any way to make these alarms trigger faster? Also, does anyone know how AWS determines scaling actions? I've attempted to use various CLI commands to uncover their scaling logic but haven't had much luck. I want to understand how to adjust scaling more effectively, especially to handle high CPU situations. My setup involves Fargate with a PHP application behind both internal and external load balancers.
1 Answer
You might want to reconsider if CPU % is the best metric to track. If your requests have consistent loads, scaling based on the request count per second from your load balancer could help you react quicker before the workers get overwhelmed. It’s worth a try!
Related Questions
How To: Running Codex CLI on Windows with Azure OpenAI
Set Wordpress Featured Image Using Javascript
How To Fix PHP Random Being The Same
Why no WebP Support with Wordpress
Replace Wordpress Cron With Linux Cron
Customize Yoast Canonical URL Programmatically