Flink: Fix JdbcLockFactory to allow ClientPoolImpl connection retry (#15759)

[yadavay-amzn]

Summary

In JdbcLockFactory.JdbcLock, the instanceId() and unlock() methods caught SQLException inside the lambda passed to ClientPoolImpl.run() and wrapped it as UncheckedSQLException (a RuntimeException). This prevented ClientPoolImpl's retry mechanism from triggering because isConnectionException() checks for SQLException, not UncheckedSQLException.

Fix

Removed the inner try-catch blocks that wrapped SQLException as UncheckedSQLException inside the lambdas. The lambda signature for ClientPoolImpl.run() already declares throws Exception, so SQLException propagates naturally to ClientPoolImpl where the retry logic can detect connection exceptions and retry appropriately.

The outer catch (SQLException e) blocks remain to convert to UncheckedSQLException after ClientPoolImpl has exhausted its retry attempts.

Closes #15759