[metrics] Add mempool monitoring series to charon-metrics exporter

[obchain]

Tracking issue filed from PR #54 dashboard review (#283).

Scope

The Grafana dashboard (deploy/grafana/charon.json) omits a mempool panel because backing series do not exist. Add:

• charon_mempool_pending_oracle_updates (gauge) — count of pending Venus oracle updates observed in the mempool
• charon_mempool_drained_total (counter) — total oracle updates drained from mempool at block confirmation, labeled by chain
• charon_mempool_websocket_reconnects_total (counter) — reconnect attempts, to catch flaky upstreams

Emission site: crates/charon-scanner/src/mempool.rs (MempoolMonitor). Wire-up tracked separately in #299.

Acceptance

• New series registered in crates/charon-metrics/src/lib.rs
• MempoolMonitor increments/sets each series on the appropriate event
• Dashboard panel added post-merge referencing these series

References

• PR feat(grafana): importable dashboard JSON for Charon metrics #54 omission list
• [cli] wire MempoolMonitor into listen loop with pre-sign drain on NewBlock #299 (CLI wire-up of MempoolMonitor)

[obchain]

Fixed on feat/22-prometheus-metrics (7258cc7): added charon_mempool_pending_oracle_updates (gauge), charon_mempool_drained_total (counter, label: chain), charon_mempool_websocket_reconnects_total (counter, label: chain). Emitted from PendingCache::insert/drain and MempoolMonitor::run reconnect branch. PendingCache and MempoolMonitor now carry the chain short-name so the label aligns with the scanner/gas-oracle stacks.