Low CPU and memory usage during high response times indicates worker starvation

warning

Resource ContentionUpdated Oct 22, 2024(via Exa)

Sources

Why your FastAPI (or Flask) App performs poorly with high loadsdev.to

Technologies:

Flasksubject

FastAPISymptoms of this issue are visible in FastAPI metrics and logs

How to detect:

When response times are high but CPU and memory usage remain extremely low, the application is not utilizing available server resources. This indicates insufficient worker processes, causing request queuing rather than parallel processing.

Recommended action:

Monitor CPU and memory utilization during load tests. If both remain low while response times degrade, increase the number of worker processes. Adjust workers based on server CPU cores and memory constraints to improve resource utilization and request concurrency.