We have a web app that is in PoC now and will soon go into production. There are 2 Cloud Run services for frontend and the backend and we want to do some capacity planning or rather understand the spend. Is it possible to get some baseline estimate for the no of cloud run instances we would need by doing some load testing something like with 200 concurrent users and 20 API calls / sec we would need 20 Cloud Run instances?
Solved! Go to Solution.
Yes, load testing is the best way to estimate your Cloud Run costs. It's easy to calculate the costs for memory and CPU etc., but the key is understanding how many instances you need to serve a certain number of requests, and that varies a lot across different services, so doing a load test like you suggest is the best way to get that number.
Yes, load testing is the best way to estimate your Cloud Run costs. It's easy to calculate the costs for memory and CPU etc., but the key is understanding how many instances you need to serve a certain number of requests, and that varies a lot across different services, so doing a load test like you suggest is the best way to get that number.
Thanks @knet . What would be a simple load testing tool Apache workbench?
Yes, capacity planning for Cloud Run is possible. It involves monitoring usage patterns, setting appropriate concurrency limits, and configuring autoscaling to manage demand effectively.