Rust Proxy Speed Optimization

I have a working Rust codebase that spins up a datacenter-grade proxy in front of a target site. Functionality is solid, but raw throughput is lagging behind expectations. The immediate goal is performance optimization with a laser focus on execution speed—latency per request must drop and overall RPS needs to climb noticeably without sacrificing stability. You’ll be diving into a fairly idiomatic async stack (Tokio + hyper) that already handles connection pooling, routing, and basic health checks. The challenge is to profile, pinpoint, and remove bottlenecks in the critical path: connection handshake, frame parsing, and response streaming. If you prefer flamegraph, cargo-bench, criterion, or any other standard Rust profiling tool, feel free to bring it in; I care about measurable results, not the specific toolchain. Deliverables • Refactored source code (PR or patch) with performance-oriented changes clearly documented • Before-and-after benchmarks showing execution-speed gains under a realistic load (wrk, vegeta, or similar) • Short report explaining what was optimized, why it helped, and any trade-offs introduced Clean, idiomatic Rust and clear commit messages will earn an immediate approval and follow-on work.