For those of you who have problems with tickets looping either from do-not-reply's [E.g. Do-not-reply notification -> ticket update notification -> do-not-reply notification -> ticket-update-notification-> etc.] or internal rules gone awry, here's some code to help alert you to the loops:

SELECT

'YOUREMAILADRESS@TEST.COM' AS TLOOP,
SUSPECT as SUSPECTT
FROM
(SELECT MAX(HD_TICKET_ID) AS SUSPECT
FROM
(SELECT  
COUNT(TIMESTAMP) AS COUNTT, HD_TICKET_ID
FROM HD_TICKET_CHANGE WHERE TIMESTAMP> DATE_SUB(NOW(), INTERVAL 1 MINUTE )
)XYT2
WHERE COUNTT>10)XYT3
WHERE SUSPECT LIKE '%%'
;

You will want to enter your own email address to recieve the alert and edit the threshold above to match your volume. If you have a very high voilume environment, its possible that you may have more than 10 interactions with tickets per minute on average.

To base this number, use the following:

-can be pasted as a SQL report in the repord tab

-Update the timeframe to match your busiest time, divede out from there to get a rate per minute to adjust the threshold above.

SELECT  
COUNT(TIMESTAMP) AS COUNTT, HD_TICKET_ID
FROM HD_TICKET_CHANGE WHERE TIMESTAMP> DATE_SUB(NOW(), INTERVAL 30 DAY )

Once you have the threshold set, setup the top rule in a test queue (disclaimer, not liable for any code, always  test before using live!!)

Set the rule to run every fifteen minutes, and to email for results row. Email column as TLOOP and body can be something like:

**Please check into into http://YOUR.KBOX.DOMAIN/userui/ticket?ID=$suspectt

Subject can be whatever grabs your attention.

Then test and adjust the threshold as needed, play with it. Suggestions for improvement welcome, but here's a start.

Nest step could be to create a rule to update the requestor in the event of a do-not-reply loop to a dummy or owner account to halt the loop, but this should at least alert you that there may be a problem to investigate.