Question

What generates the PEGA0017 alert and how can it be addressed?

Response

The PEGA0017 alert is triggered when a cache size exceeds the limit value and indicates that the instances of the stated cache are being invalidated (purged).

Like PEGA0016, this setting relates to various caches in the system. When entries stored in a cache exceed the limit value (not the target value) that was set in the prconfig.xml file (see Default prconfig.xml settings below), the oldest entries in the cache are immediately invalidated until the count of entries in the cache is below the target limit. The alert message specifies which cache was affected:

• fua/global
• fua/personal
• PropertyReference
• LowerCase
• UpperCase

This alert is available in PRPC 5.1 and later versions.

Alert message description

Cache forcibly reduced to limit size.

Example of message text

collections/mru/UpperCase: Successfully reduced MRU by 3002 down to: 15000.

Default prconfig.xml settings

<env name="fua/global/instancecountlimit" value="20000" />

<env name="fua/personal/instancecountlimit" value="20000" />

<env name="collections/mru/PropertyReference/instancecountlimit" value="10000" />

<env name="collections/mru/LowerCase/instancecountlimit" value="10000" />

<env name="collections/mru/UpperCase/instancecountlimit" value="15000" />

Probable reasons for alert and next steps

When this alert occurs, it is a strong indication that the cache limit in the prconfig.xml file is too low.

A cache that is too small hurts performance by requiring more objects to be fetched from the database (or fetched from the file system or loaded by a class loader) rather than from the cache. (However, a cache that is too large can also hurt performance, because it reduces the memory that is available for other processing within the JVM.)

Calculate the optimal setting for the caches and adjust the prconfig.xml settings accordingly.