Add --daemon-rpc-fallback-addr for backup bitcoind RPC

[EddieHouston]

Summary

• Adds a new --daemon-rpc-fallback-addr CLI flag (optional) that specifies a backup bitcoind JSON-RPC endpoint.
• When the primary --daemon-rpc-addr is unreachable, electrs immediately tries the fallback before entering the 3-second backoff loop.
• On every reconnect the primary is attempted first, so electrs automatically migrates back once the primary recovers.

Motivation

Allows running electrs against a pair of bitcoind nodes (e.g. primary + warm standby) so that transient restarts or maintenance on the primary don't stall indexing.

Changes

• config.rs — new daemon_rpc_fallback_addr: Option<SocketAddr> field and CLI arg
• daemon.rs — tcp_connect tries primary then fallback; Connection stores and propagates the fallback through reconnect()
• electrs.rs, tx-fingerprint-stats.rs — pass the new field through to Daemon::new

Test plan

• Start electrs with only --daemon-rpc-addr (no fallback) — verify behavior is unchanged
• Start electrs with --daemon-rpc-fallback-addr pointing to a second bitcoind, stop the primary — verify electrs connects to fallback
• Restart the primary — verify electrs reconnects to primary on next reconnect cycle