I'm part of a startup, and managing AWS costs is crucial for us—these expenses can flip a month from being profitable to a loss. We use EC2 instances (mostly t3.medium and m5.large) and rely on Lambda for data processing. Our monthly bills vary widely, ranging from $2,000 to $10,000 due to demand spikes. We've tried turning off unused instances and monitoring with CloudWatch, but our spending is still out of control. Recently, we started using a tool called Milkstraw, similar to PUMP, which has improved things a bit, but I'm looking for more strategies or optimization tips to lower these costs. We're super bootstrapped, so hiring a full-time expert isn't an option right now.
1 Answer
First off, can you share a bit about your architecture? Do you have autoscaling set up? Understanding your usage spikes—like what's causing them—could really help. Have you thought about using spot instances or Savings Plans? Those are solid starting points for cutting costs. A deeper analysis of your software and infrastructure could also reveal opportunities for savings. Let me know if you're comfortable sharing more info!
We mainly get spikes when users launch marketing campaigns or onboard strong numbers from free tiers. Autoscaling is in place, and we’re working with Milkstraw to improve that further. I appreciate the help!