Thread.sleep is one 1000 times too fast

[olavloite]

The example usage for TransactionManager contains an error in the retry part: The returned timeout is already in milliseconds, but the example divides it by 1000 before passing it to Thread.sleep.

[elharo]

Moved this to P1 since it's not just in the docs or the tests.

[olavloite]

Moved this to P1 since it's not just in the docs or the tests.

For clarity and the release notes:
The call that used the wrong value that is not in documentation or tests, is only for the JDBC driver. So the actual bug (other than wrong documentation) does not affect the client library, but the JDBC driver.