Phase 12: Add observability, diagnostics and performance benchmarks

[ashishsinghbora]

Goal

Make performance and failures measurable instead of relying on guesses.

Scope

• Add structured logging.
• Add metrics for bytes transferred, active workers, retries, failures, throughput and queue depth.
• Add optional debug diagnostics.
• Add benchmark suite for single-stream vs ranged downloads.
• Benchmark different chunk sizes and worker counts.
• Test memory usage on large files.
• Add race-detector CI runs.
• Document how contributors reproduce performance results.

Acceptance criteria

• go test -race ./... passes.
• Benchmarks are reproducible with a local controlled HTTP server.
• Logs contain enough context to diagnose a failed job without leaking secrets.
• Performance changes can be compared before/after.