[BUG] nostream info hangs indefinitely when the database is unreachable

[Priyanshubhartistm]

Describe the bug

nostream info (and nostream info --json) hangs forever instead of exiting or printing an error when the configured database is unreachable wrong host, DB down, network partition, etc.

The root cause is in getEventCount() (src/cli/commands/info.ts): it opens a knex connection pool for a single ad-hoc query. When the initial connection attempt itself fails or times out, the pool's destroy() call resolves without actually closing the underlying socket the in-flight connection attempt is left dangling, which keeps at least one open handle alive and prevents the Node process from ever exiting on its own.

To Reproduce

1. Point the CLI at an unreachable/misconfigured database, e.g. an invalid DB_HOST or a DB that isn't running.
2. Run nostream info or nostream info --json.
3. The process never returns — no error, no timeout, no exit. It has to be killed manually

Confirmed directly via process._getActiveHandles(): after the pool's destroy() resolves, 5 leaked Socket handles are still present, which is what keeps the event loop alive.

Expected behavior

nostream info should either return a result or fail with a clear error/timeout within a bounded time, the same way every other CLI command already behaves when the database is unavailable it should never hang indefinitely.

System (please complete the following information)

• OS: Linux
• Platform: standalone / Docker Compose (reproduces in both the CLI's own DB access, not the
  relay server)
• Version: current main