`evaluation/runQueries` query server command fails silently if it fails to acquire the database lock

Hi!

When having several query servers running on the same database (eg: VSCode + CodeQL.nvim) the first one running a query acquires a lock for the database, and running a query using `evaluation/runQueries` command on the second one fails silently without returning the reason of the failure.

Start at the evaluation/runQueries query server command and trace how database-lock acquisition failures are handled. Reproduce the case with two query servers using the same database; done means the second command returns the reason it could not acquire the lock instead of failing silently.

Written by the indexing model from the issue text.