fix: eventbus index removal, ordered channel query logging, and dst chain log

[hitchhooker]

Summary

fixes three bugs that could cause events to be lost or make debugging harder.

Changes

1. EventBus: fix index removal when multiple subscribers disconnect

File: crates/relayer/src/event/bus.rs

When multiple subscribers disconnected simultaneously, removing indices in ascending order caused off-by-one errors:

// Before: indices [1, 3] - removing 1 shifts 3 to 2, then removing 3 removes wrong element
for idx in disconnected {
    self.txs.remove(idx);
}

// After: remove in reverse order to preserve correct indices
for idx in disconnected.into_iter().rev() {
    self.txs.remove(idx);
}

Added test to verify correct behavior with multiple disconnected subscribers.

2. Packet worker: log errors when querying next_sequence_receive fails

File: crates/relayer/src/worker/packet.rs

Previously errors were silently swallowed with .ok(). Now logs a warning so operators can see query failures while preserving the fallback behavior (triggering packet clearing).

3. Supervisor: fix wrong chain id in log message

File: crates/relayer/src/supervisor.rs

Was logging src_chain_id when dst_chain_id failed to spawn. Now correctly logs dst_chain_id for easier debugging.

Test Plan

• cargo check -p ibc-relayer passes
• cargo test -p ibc-relayer --lib passes (73 tests, +1 new)
• new test multiple_disconnected_subscribers verifies EventBus fix