Incoming request limit from another API


We have a problem with a new API we try to deploy on Render. We do deploy our API on GCP Cloud Run until now, and we want to deploy our new API on Render to test the service.

Our problem is that our main API on GCP gets stuck calling the API on Render.
The API on Render is called by clients without any problems (2/s from clients) but the calls from our main API (15/s) are stuck a few seconds after the main API is launched.

We tried to “just” return “ok” from the API on Render and to force a minimal count of 3 instances but it changed nothing. What is strange is that the CPU and memory usage are very low when we have around 3 minutes of response time while sending “ok”.

We are a bit lost, and we tried, in a last resort, to call a url instead of our new API and the average response time went back to normal (~40ms).

As it “work” with, we think we get rate limited in some way by Render (or your cloudflare ?). Can anybody help ?