Ceph

When we tried to solve the problem of slow requests in the new cluster, we found that the official website documentation had very little description and was out of date. Can anyone help update the documentation?
Link is here:
https://docs.ceph.com/en/latest/rados/troubleshooting/troubleshooting-osd/#debugging-slow-requests

By the way, can you tell where the performance bottleneck is from the attached information?
We use 8+2 ec to deploy the ceph cluster, 10 nodes and 21 nvme disks per node. The test uses 4MB objects to stress test 10 rgw gateways.