RPC
Staging uses publicnode for every process. WS and HTTP must be the same provider.
https://gnosis-rpc.publicnode.com
wss://gnosis-rpc.publicnode.comMixing providers (WS on A, HTTP on B) produces w3: call failed: not found on some blocks: the websocket announces a head the HTTP node does not have yet.
eth-tracker batches up to 100 JSON-RPC sub-requests per call. Default pool_size = 0 is runtime.NumCPU()*3. Staging sets pool_size = 4 in tracker-config.toml.
Self-hosted
Run gnosis-node (reth + nimbus) and point:
TRACKER_CHAIN__RPC_ENDPOINT=http://127.0.0.1:8545
TRACKER_CHAIN__WS_ENDPOINT=ws://127.0.0.1:8546
CORE_CHAIN__RPC_ENDPOINT=http://127.0.0.1:8545
INDEXER_CHAIN__RPC_ENDPOINT=http://127.0.0.1:8545Archive (or a node that has logs from block 48027888) is required for tracker catchup.