Resetting Alerts and Test Result Evaluations in Mailscape Database

Overview

Occasionally the Alerts table in the Mailscape database enters a bad state, producing multiple alerts for the same alert category and server combination. If that happens, you can reset all alerts and force the data models to be re-evaluated.

Symptoms

• Multiple alerts firing for identical alert category and server combinations.
• Alert actions not behaving as expected in Mailscape monitoring.

Before you run this

These statements are destructive and global. They do not target the affected server or alert category — they clear the evaluation table entirely and delete every current alert in the system, for every monitored server.

• Back up the Mailscape database first. TRUNCATE TABLE cannot be undone once committed.
• Confirm which database you are connected to before executing. These are unqualified table names, so they act on whichever database the query window is pointed at.
• Run this during a maintenance window if you can. Between the reset and the next evaluation cycle, current alert state is gone.

Resolution

1. Connect to the Mailscape database using SQL Server Management Studio or your preferred query tool.

2. Run the following:

   TRUNCATE TABLE [dbo].[TestResultEvaluations]
   DELETE FROM [dbo].[Alerts]

   • TRUNCATE TABLE on [TestResultEvaluations] removes all rows, resetting evaluation data.
   • DELETE FROM on [Alerts] removes all current alerts from the system.

Run the two statements together and check that both succeeded. They need different permissions — TRUNCATE requires ALTER on the table while DELETE does not — so it is possible for the second to succeed while the first fails, which leaves the reset half-finished.

What happens next

• All current alerts are cleared.
• The system re-evaluates all models. Anything that still qualifies for an alert generates a new one on the next cycle, so expect legitimate alerts to return shortly — that is the reset working, not the problem recurring.