Within 1.3 build 3, the retry mechanism of Outscaling for sending failed events has been changed slightly. However, these changes has introduced a bug, which has to do with comparing a NULL value against a numeric value on the RetryCount field of the osEventQueueStatus table.

The result is that when sending an event fails for whatever reason, this event is not marked as being failed and the error that occurred won't be written to the EventQueue Errors (osEventQueueStatus table as well). Because of this, the Outscaling background component will endlessly keep trying to send this event (until it succeeds) every background loop.

But some events will never succeed (e.g. due to database constraint violations), so these events are the problem. Outscaling will keep trying to send these events, and for the sake of performance and resource usage this should be prevented.

Outscaling will be fixed within the next build/release (isnull check added to the retrycount comparing), however in the meantime it is strongly adviced to add a database DEFAULT CONSTRAINT (with value 0) to the RetryCount column of the osEventQueueStatus table. You can use the SQL script shown below.

SQL	Copy Code
IF NOT EXISTS (       SELECT 1 FROM sys.default_constraints  dc       INNER JOIN sys.columns c ON dc.parent_object_id = c.object_id AND dc.parent_column_id = c.column_id       WHERE parent_object_id = object_id('osEventQueueStatus')       AND c.name = 'retrycount' ) ALTER TABLE osEventQueueStatus ADD CONSTRAINT DF_osEQStatus_RetryCount DEFAULT ((0)) FOR retrycount;

This will, in the meantime, effectively fix the bug mentioned above and will have no impact once you install the next build/release containing the bug-fix in code.