AEM Cloud Adaptive Rate Limiter: Understanding “Server Load Exhausted” and Prediction

Hello,

We are investigating an issue in AEM as a Cloud Service where some customer requests are being rejected by the Adaptive Rate Limiter with the message:

server load exhausted

The logs show that the rejection is based on the calculated server load and the predicted additional load, rather than the maximum number of in-flight requests being reached.

*INFO* [qtp1005821649-5497] com.adobe.granite.ratelimit.AdaptiveRateLimitFilter Rejected <url>, server load exhausted. Server load: 749009/750000, inflight: 68/200, prediction: 35381

Any guidance on interpreting these Adaptive Rate Limiter log fields and recommended troubleshooting steps would be appreciated.

Thanks,
Prince