[PR #32] Exponential backoff has no jitter — thundering herd on BSC WebSocket provider

[obchain]

PR: #32 feat/07-block-listener
File: crates/charon-scanner/src/listener.rs — BlockListener::run loop

Problem

Reconnect loop doubles backoff on every failure with no random jitter:

backoff = (backoff * 2).min(Duration::from_secs(30));

When BSC WS provider restarts or rate-limits (common on public endpoints), every listener task that disconnected at same instant retries at identical intervals: 2s, 4s, 8s, … 30s. Requests arrive in lockstep, overwhelming provider and triggering another rejection wave — thundering-herd loop.

Impact

All chain listeners share same WS endpoint config. Provider blip that disconnects all listeners simultaneously causes correlated retry storms keeping bot offline far longer than underlying outage.

Fix

Add 0-25% random jitter before sleeping:

use rand::Rng;
let jitter = rand::thread_rng().gen_range(0..=backoff.as_millis() / 4);
tokio::time::sleep(backoff + Duration::from_millis(jitter as u64)).await;
backoff = (backoff * 2).min(Duration::from_secs(30));

Add rand to workspace deps. 30s cap is correct; only jitter missing.

[obchain]

Fixed on feat/07-block-listener in a802e2a.