How to troubleshoot gRPC request timeouts between Zeebe Broker and Gateway?

I’m using Zeebe version 8.3.0. I execute approximately 5000+ process instances per second. To achieve faster execution of the processes, I have deployed 27 Brokers without any replicas. However, I have encountered many errors in the service, as shown in the image below. To troubleshoot the issue, I improved the monitoring and found that the gateway latency is as high as 10 seconds. 26% of the instances are timing out, and the minimum execution interval between two nodes in the process is at least 1 second. I don’t understand why, with 27 Brokers, I am unable to achieve a throughput of 5000. Can someone help me investigate this problem?

Is there any issue with the number of gRPC invocations?

Does anyone know about it