We can definitely add an info level message for it but a notice level message would be too verbose and would not really bring any relevant information for the majority of users.

The biggest problem is that the way the database refreshing is implemented: When it is time to refresh the database list, all active users of that list will try to update it. This means that every query that arrives after the refresh has started will trigger a new round of database queries. This pessimistic approach would work if the rate of failure for updates would be very high. As the listing of databases is almost guaranteed to succeed, this just causes an extra overhead in updating of the lists.

The correct way to do the update would be to assign it to one connection and re-assing it if it fails.

I agree, Markus. Maybe having only one connection doing it and making it as done in case the last iteration to create refresh the mapping list returns true would be enough. The intention is to give it more verbosity when the database mapping list is updated. I understood the design limitation also. Having this information will help to detect possible connection denies, which in this scenario, can be interpreted as a connection that tries to use a non-existing database (non-existing on the mapping list, but physically created)... right?

We'll add an info level message so that at least some trace is left in the log when the database map is updated.