AES considers the PEGA0044 alert (alerts throttled) to be a critical event. Since Pega7 processes exception handling through the alert processor, when there are a lot of exceptions (300 in 5 minutes by default) it triggers the PEGA0044 and AES triggers both the 'throttle' scorecard and the 'health status change' scorecard.
Best thing to do of course is to fix the exceptions. I personally feel that a healthy system should never generate exceptions, though the Connect-SOAP implementation treats all SOAP faults as exceptions, regardless of whether they are from true failures or simply a mechanism for a back-end to say "no" (btw, if this is the cause of your exceptions, it is possible to change a rule and change this behavior).
Barring making the exceptions go away consider to change the PEGA0044 threshold - the defaults are pretty low. If you read the thread, prconfig settings were posted by sriva3. I personally consider editing prconfig.xml to be a bad practice and use dynamic system settings. In my systems, we only throttle / discard alerts if you get more than 5,000 in 5 minutes [system is truly broken]