[EPIC] BlockSync Overhaul - Networking

[Wondertan]

Context

This EPIC is an entry point for the BlockSync Networking Overhaul efforts and a continuation of #1099
For deeper context, look at the ADR document: #1326.

Tasks

• Merge docs(adr-13): Block Data Sync Overhaul: Part II - Network #1326

New APIs

ShrEx/EDS

• eds: Client #1416
• eds: Server #1417
• modshare: Provide eds.Client and eds.Server to FX #1418

ShrEx/ND

• eds: NDClient #1419
• eds: NDServer #1420
• modshare: Provide eds.NDClient and eds.NDServer to FX #1421

ShrEx/Push

• eds: PubSub #1432
• modshare: Provide eds.PubSub to FX #1433

share.Getter

• share/get: blockservice.BlockGetter adapter for eds.Store #1563
• share: Getter interface and IPLDGetter #1488
• feat(share/peer manager): On sync broadcast new message instead of retransmitting old one. #1692
• share/get: ShrExGetter #1489
• share/get: StoreGetter #1490
• share/get: CascadeGetter #1491
• node/share: Integrate share.Getter #1492
• share/get: TeeGetter #1535
• share: IPLDGetter should use cachelss only BlockGetter #1666

Integrations

Discovery

• discovery: Add Peers method #1449

Availability

• share: Extend Availability interface with peering info #1450
• availability/full: Implement happy path EDS retrieval via ShrEx/EDS #1464
• node/share: Data sync modes for FN #1465

Listener

• core: Broadcast ShrEx/Sub notification in Listener #1467

DASer

• das: Integrate ShrEx/Sub using eds.PubSub #1466

Testing

• swamp: GetEDS test using polling for FN and LN
• swamp: GetEDS test using shrex/sub FN and LN
• testground: GetEDS tests over API

Hardenings

• architecture (share/Getter): Getter should be atomic #1663
• bug(nodebuilder/share): providing shrex.Getter and its dependencies cause unexpected error in first sampling attempt #1771
• bug(share/p2p/shrexsub | das): don't broadcast empty blocks' data hashes #1775
• feat(share/shrex): add Rate limit middleware for shrex servers #1745
• share/p2p: ShrexSub: track msg duplicates via pubsub.Tracer #2745
• share/getter: Move to paralle/async CascadeGetter #1635
• feat(share/shrex/peer_manager): add feedback option to put peer on cooldown #1731
• feat(share/shrex/peer_manager): add debug logs to peer manager #1738)
• share/p2p/peers: add metrics to peer manager #1780
• feat(share/p2p/peer-manager): optimise peer manager memory footprint #1781

Reconstruction

• eds: Retriever requests all the quadrants at once, instead of one-by-one

Bugs

• ShrEx bugs #1779