[zlacfzy]

Description

eth: tune timing parameters and broadcast queue for 450ms block

Rationale

• Reduce block/tx fetcher arrive timeouts from 500ms to 200ms and gather slack from 100ms to 50ms, so the effective fetch trigger (150ms) stays well within the 450ms block interval. The previous 400ms trigger could
  cause announce-path blocks to arrive after the next block was already due, leading to cascading propagation delays.
• Increase per-peer block broadcast queue from 4 to 16 to absorb bursts when occasional slow blocks (e.g. 1.5s execution) temporarily stall consumption, preventing dropped block broadcasts that force slower recovery via announce+fetch or sync.

Example

add an example CLI or API response...

Changes

Notable changes:

• add each change in a bullet point here
• ...