Add "keep alive" mechanism

It would be nice to have a "keep alive" mechanism that periodically checks if the connection(s) to the defined host(s) is/are active and through an exception after a timeout or an x times timeout and disconnect.

This way a user won't potentially wait for "infinite" time, for a long running query to return, if in between the connection(s) is completely lost (cluster is down, etc.).

Start by tracing the CLI's query execution and connection-handling entry points; the issue does not name files or tests. Clarify the timeout, retry-count, exception, and disconnect behavior, then define completion around detecting lost connections during long-running queries without waiting indefinitely.

Written by the indexing model from the issue text.