Inconsistent log level used in liveness checkers

[opolsen]

The log level used in the IsHealthy method seems to be inconsistent across the various liveness checkers.

There might be some reasoning behind this, but if not then I think it would be preferable to decide on which log level to use and apply that to all the liveness checkers.

Some examples:
DynamoDb, NpgSql: Logs both success and exception as debug
Oracle: Logs success as debug and exception as warning
MySql, SqlServer: Logs success as information and exception as warning

[unaizorrilla]

I try to work on this ASAP