Scale up the pod number of bookie or broker cant improve the performance of throughput and my hardware resource is still free enough

**Describe the bug**
I deployed pulsar with k8s. We need 25w/s(1k size msg) throughput, however it can only provide 9w/s throughput with perf tool. By the way, its far away from the official test report. I tried to scale up the pod number of bookie and broker, but it worked nothing for me and I found my hardware resource is still free enough. What's wrong with my operation?

**To Reproduce**
See description

**Expected behavior**
Support horizontal expansion

**Screenshots**
If applicable, add screenshots to help explain your problem.

**Desktop (please complete the following information):**
- OS: k8s, pulsar version 2.8

**Additional context**
Add any other context about the problem here.

Start with the Kubernetes deployment, Pulsar 2.8 configuration, and the perf tool setup described in the issue. Compare throughput before and after changing bookie and broker pod counts, while recording resource usage and the benchmark parameters. Done means identifying why horizontal expansion does not improve the reported throughput or documenting the missing configuration or reproduction details.

Written by the indexing model from the issue text.