[ChainOrchestrator] Refactoring / Improvements

[frisitano]

Overview

This issue captures improvements/modifications to the ChainOrchestrator, which was introduced in #185.

Issues

• Handling L1 provider unavailability and signaling optimistic sync #252
• [ChainOrchestrator] Standardise data structures for received_chain_headers and current_chain_headers #268
• L1 live sync message validation (pre Euclid V2) #245

Refactor

• Leverage GetBlockHeaders to fetch a batch of headers at once - this is more efficient than fetching one by one
• Instead of maintaining an in-memory buffer, we should leverage the execution node more heavily.
• Consider reorg condition - currently we use timestamp to determine the most recent fork - is this suitable?