`Cluster` database engine

A follow-up to the `Remote` database engine (https://github.com/ClickHouse/ClickHouse/issues/59304, implemented in https://github.com/ClickHouse/ClickHouse/pull/110975).

Implement a new database engine, `Cluster`. While the `Remote` database engine takes explicit addresses, the `Cluster` engine should take the name of a cluster from the server configuration — the same way the [`cluster`/`clusterAllReplicas`](https://clickhouse.com/docs/sql-reference/table-functions/cluster) table functions relate to the [`remote`](https://clickhouse.com/docs/sql-reference/table-functions/remote) table function:

```sql
CREATE DATABASE db ENGINE = Cluster('cluster_name', 'database');
```

It provides real-time access to the tables of a database on the cluster: the list of tables and their structure are fetched on demand, and every table is exposed as a `Distributed` storage over the named cluster, forwarding `SELECT` and `INSERT` queries to it.

Related: https://github.com/ClickHouse/ClickHouse/issues/59304
Related: https://github.com/ClickHouse/ClickHouse/pull/110975

Start by reading issue 59304 and pull request 110975, which describe the related Remote database engine. Compare that behavior with the requested CREATE DATABASE db ENGINE = Cluster('cluster_name', 'database') semantics; done means the cluster-backed database exposes current table structure and forwards SELECT and INSERT queries through the named cluster.

Written by the indexing model from the issue text.