Native prices: Implement generic buffered batching system 1/3 (#2869)

# Description
Implements a generic batching system for the native prices.

The implementation has a blocking function
`blocking_buffered_estimate_prices()` that belongs to a struct which
runs a thread. The thread gathers the requests either by time (it waits
X time to gather request and then fires them) or by number of requests
(as soon ass it reaches a number of requests, fires them).

This generic implementation allow for any native price fetcher to batch
buffer the request as long as they implement the trait
`NativePriceBatchFetcher`.

This code is heavily based on the `ethrpc` crate code.

In the upcoming PR:
- the configuration will be propagated to the services configuration
- implement the trait `NativePriceBatchFetcher` for CoinGecko

# Changes
- Generic buffered batching system for native prices

## How to test
1. Unit test

---------

Co-authored-by: MartinquaXD <[email protected]>

crates/shared/src/price_estimation.rs

@@ -26,6 +26,7 @@ use {
     thiserror::Error,
 };
 
+mod buffered;
 pub mod competition;
 pub mod external;
 pub mod factory;