I Got 100% usage of Fluid Compute CPU, yes this is original from our backend infrastructure, we using ExpressJs for APIs inference and 3 other database, the frequently executed is from the postgress DB about ~400 executed everyday. IDK how the Fluid compute cpu works and whats the caused problem from this infra. Is the uptime bot still count the cpu to be always online? and is the redirect caused the cpu always on too? or its pure my traffic? here my backend url https://api.neosantara.xyz and this the running dashboard using backend url api neosantara https://app.neosantara.xyz this processes fetch user detail like username, logs etc. also when calling api neosantara there have an logs processing like tier user, total consume tokens etc..
Looking at your situation, Fluid Compute CPU usage appears to be based on actual compute time when your functions are actively processing requests, not just uptime or redirects.
Main factors that could be contributing to your high CPU usage:
- Database queries: 400+ PostgreSQL queries daily can add up, especially if they’re complex or unoptimized
- API processing: Your Express.js inference APIs likely consume significant CPU during processing
- Concurrent requests: Multiple simultaneous API calls can spike CPU usage
What doesn’t count toward CPU:
- Uptime monitoring bots (they just ping endpoints)
- Simple redirects
- Static file serving
To optimize:
- Review your database queries for efficiency
- Implement caching where possible
- Consider connection pooling for your PostgreSQL connections
- Monitor which specific API endpoints consume the most resources
You can track your usage in the Vercel Dashboard under Usage > Functions to see which functions are consuming the most CPU time.
1 Like
how can i pause my project? so the compute cpu can be inactive temporary then when cpu is cold i can online again. Can i do that?
Spend Management would let you automatically pause your project on a Pro plan to avoid a runaway bill. You cannot manually pause a project on the Hobby plan. Instead, you can optimize to reduce usage or remove the project if it is no longer needed.