feat: active filter subscription management for edge nodes
## Subscription Manager
* edgeFilterSubLoop reconciles desired vs actual filter subscriptions
* edgeFilterHealthLoop pings filter peers, evicts stale ones
* EdgeFilterSubState per-shard tracking of confirmed peers and health
* best-effort unsubscribe on peer removal
* RequestEdgeShardHealth and RequestEdgeFilterPeerCount broker providers
## WakuNode
* Remove old edge health loop (loopEdgeHealth, edgeHealthEvent, calculateEdgeTopicHealth)
* Register MessageSeenEvent push handler on filter client during start
* startDeliveryService now returns `Result[void, string]` and propagates errors
## Health Monitor
* getFilterClientHealth queries RequestEdgeFilterPeerCount via broker
* Shard/content health providers fall back to RequestEdgeShardHealth when relay inactive
* Listen to EventShardTopicHealthChange for health recalculation
* Add missing return p.notReady() on failed edge filter peer count request
* HealthyThreshold constant moved to `connection_status.nim`
## Broker types
* RequestEdgeShardHealth, RequestEdgeFilterPeerCount request types
* EventShardTopicHealthChange event type
## Filter Client
* Add timeout parameter to ping proc
## Tests
* Health monitor event tests with per-node lockNewGlobalBrokerContext
* Edge (light client) health update test
* Edge health driven by confirmed filter subscriptions test
* API subscription tests: sub/receive, failover, peer replacement
Co-authored-by: Ivan FB <128452529+Ivansete-status@users.noreply.github.com>
Co-authored by Zoltan Nagy
* Fix peer selection for cases where ENR is not yet advertiesed but metadata exchange already adjusted supported shards. Fix initialization rendezvous protocol with configured and autoshards to let connect to relay nodes without having a valid subscribed shard already. This solves issue for autoshard nodes to connect ahead of subscribing.
* Extend peer selection, rendezvous and metadata tests
* Fix rendezvous test, fix metadata test failing due wrong setup, added it into all_tests
* Adapt using chronos' TokenBucket. Removed TokenBucket and test. bump nim-chronos -> nim-libp2p/nim-lsquic/nim-jwt -> adapt to latest libp2p changes
* Fix libp2p/utility reports unlisted exception can occure from close of socket in waitForService - -d:ssl compile flag caused it
* Adapt request_limiter to new chronos' TokenBucket replenish algorithm to keep original intent of use
* Fix filter dos protection test
* Fix peer manager tests due change caused by new libp2p
* Adjust store test rate limit to eliminate CI test flakyness of timing
* Adjust store test rate limit to eliminate CI test flakyness of timing - lightpush/legacy_lightpush/filter
* Rework filter dos protection test to avoid CI crazy timing causing flakyness in test results compared to local runs
* Rework lightpush dos protection test to avoid CI crazy timing causing flakyness in test results compared to local runs
* Rework lightpush and legacy lightpush rate limit tests to eliminate timing effect in CI that cause longer awaits thus result in minting new tokens unlike local runs
* fix: admin API peer shards field from metadata protocol
Store and return peer shard info from metadata protocol exchange instead of only checking ENR records.
* peer_manager set shard info and extend rest test to validate it
Co-authored-by: MorganaFuture <andrewmochalskyi@gmail.com>
* properly pass userMessageLimit to OnchainGroupManager
* waku.nimble 2.2.4 Nim compiler
* rm stew/shims/net import
* change ValidIpAddress.init with parseIpAddress
* fix serialize for zerokit
* group_manager: separate if statements
* protocol_types: add encode UInt32 with zeros up to 32 bytes
* windows build: skip libunwind build and rm libunwind.a inlcusion step
* bump nph to overcome the compilation issues with 2.2.x
* bump nim-libp2p to v1.10.1
* Extended /admin/v1 RESP API with different option to look at current connected/relay/mesh state of the node
* Added score information for peer info retrievals
* add new unit test to validate that any peer can be retrieved
* add new discv5 test and better peer store management
* wakuPeerStore -> switch.peerStore
* simplify waku_peer_store, better logs and peer_manager enhancements
Better control when the remote peer closes the WakuFilterPushCodec
stream.
For example, go-waku closes the stream for every received message.
On the other hand, js-waku keeps the stream opened.
Therefore, we support both scenarios.
* waku/waku_filter_v2/protocol.nim keeps track of the filter-client connections in Table[PeerId, Connection]
* waku/waku_filter_v2/protocol.nim starts listening for peer-left events in order to completely remove the previous Connection instance. Also, a new Connection is added when the filter-service starts publishing to its peers.
---------
Co-authored-by: NagyZoltanPeter <113987313+NagyZoltanPeter@users.noreply.github.com>
* bump_dependencies.md: add nim-results dependency
* change imports stew/results to results
* switching to Nim 2.0.8
* waku.nimble: reflect the requirement nim 1.6.0 to 2.0.8
Adding --mm:refc as nim 2.0 enables a new garbage collector that we're
not yet ready to support
* adapt waku code to Nim 2.0
* gcsafe adaptations because Nim 2.0 is more strict
