op-conductor flashblocks-streaming and raft timeout issues

Identify the root cause behind the issue of conductor trying to forward flashblocks from rollup-boost.

One of the identified issues has been

```
"error reading from rollup boost WebSocket" err="failed to read: read limited at 32769 bytes"
```

Moreover, there had been sporadic issue around conductor raft p2p timeout expiring causing it to transfer leadership despite no unhealthiness.

[Example]([https://optimistic.grafana.net/a/grafana-lokiexplore-app/explore/service/op-conductor/lo[…]sage=false&sortOrder=%22Ascending%22&wrapLogMessage=false](https://optimistic.grafana.net/a/grafana-lokiexplore-app/explore/service/op-conductor/logs?patterns=%5B%5D&from=2025-10-12T01:57:00.000Z&to=2025-10-12T01:59:00.000Z&var-lineFormat=&var-ds=grafanacloud-logs&var-filters=service_name%7C%3D%7Cop-conductor&var-filters=network%7C%3D%7Cmainnet-prod&var-filters=app_kubernetes_io_instance%7C%3D%7Cprod-mainnet-op-opn-geth-f-sequencer-1&var-fields=&var-levels=&var-metadata=&var-jsonFields=&var-patterns=&var-lineFilterV2=&var-lineFilters=&timezone=browser&var-all-fields=&displayedFields=%5B%5D&urlColumns=%5B%5D&visualizationType=%22logs%22&prettifyLogMessage=false&sortOrder=%22Ascending%22&wrapLogMessage=false))