Summary

• Keep the original --timeout deadline across all retry attempts.
• Cap retry sleeps to the remaining budget and fail before the next attempt when the budget is exhausted.
• Refresh the per-request reqwest timeout before every send so retries inherit the current remaining time.
• Add a best-effort time cap to response draining on retry paths.
• Add integration coverage for timeout exhaustion on both retryable status responses and retryable transport errors.

Testing

• cargo fmt
• cargo clippy --locked --all-targets --all-features -- -D warnings
• cargo test --all-features --test integration -- --test-threads=1
• cargo test --all-features --bin fetch