Is your feature request related to a problem? Please describe

Investigate GRPC server optimizations for higher throughput.

Current Issues

1. Event Loop Blocking: gRPC service methods execute directly on Netty event loop threads via directExecutor(), causing I/O blocking during protobuf serialization
2. Limited Concurrency Control: No dedicated setting to tune gRPC request processing parallelism independently of network I/O threads
3. Shared Event Loop Groups: GRPC boss and worker event loops are not properly separated, leading to connection acceptance competing with I/O processing

Describe the solution you'd like

1: Dedicated gRPC Executor Thread Pool (solves #1 and #2)

• Replace directExecutor() with a dedicated ForkJoinPool for processing gRPC service calls
• Add grpc.netty.executor_count setting to control this GRPC request processing thread pool size
• (Note this is in contrast with grpc.netty.worker_count which is controlling the number of network I/O threads)

2: Separated Event Loop Groups (solves #3)

• Prevent connection acceptance from blocking I/O processing:
  • Boss Event Loop Group: Dedicated to accepting new client connections
  • Worker Event Loop Group: Dedicated to network I/O operations (grpc.netty.worker_count threads)

Related component

Search:Performance

Describe alternatives you've considered

No response

Additional context

No response