How Declarative Cache count is calculated in Memory Management
Hi All,
We got Pega 17 alert and through DSS setting we updated it to 5000. For couple of months, it was fine, but later we started again getting Pega 17 alerts.
when we checked across all the production nodes, we see that the Declarative Cache count under memory Management is 6800+ approximately.
Now we can set the DSS setting to 10000, But this is kinda of like suppressing the alert each time, it reaches the threshold.
We wanted documentation\reference materials on how the Declarative Cache count and other cache count is calculated and displayed under Memory Management. Also Pega must be using some logic to identify and throw each and every alert.
We raised an SR with GCS, asking for any documentation\reference material available for all the Pega alerts. This will help us understand the root cause of each alert better and take preventative steps during development phase itself.
However, GCS closed the SR and advised to post a query in Community.
If any one has the logic details on one or more of the Pega alerts, please kindly share.
***Edited by Moderator: Pallavi to update SR Details***
Pegasystems Inc.
IN
Hi,
This pega0017 alert occurs when certain cache sizes configured are insufficient for the environment. As a result of undersized caches, the system is forced to purge items from the cache regardless of usage. This can also cause an unusually high number of cache misses and negatively impact system performance.
Can you please create the below DSSs with the following Setting Purposes and Values (modify the value if the DSSs are already exist).
Purpose: prconfig/collections/mru/propertyreference/instancecountlimit/default Value: 200000
Purpose: prconfig/declaratives/fragments/instancecountlimit/default Value: 50000
Note that these values need to be raised based on your environment requirements, but larger caches impact the overall JVM memory usage.
https://community.pega.com/support/support-articles/high-number-pega0017-cache-reduced-alerts
Thank You,
ANZ Banking Group Ltd
IN
Hi Nandita,
Thanks for your response.
The alert we have got is *declaratives/fragments: Successfully reduced Concurrent MRU Map by 1001 down to: 5000.*
Our current DSS setting is Pega-Engine • prconfig/declaratives/fragments/instancecountlimit/default = 5000
Increasing it to a higher value say 10000 or 100000 should solve the issue.
But what we are looking for is how is the Declarative cache count is calculated?
By running a query direcly in DB, I was able to find 4800+ Unique Declarative rules. (select distinct pxinsname from pr4_rule_vw where pyclass like 'Rule-Declare-%').
Now not all the rules will get invoked and landed up in the cache.
Even if it does so, it is still less than 5000.
But why we are seeing 6800+ as the Declarative cache count? How the SMA calculates and displays the cache counts under Memory Management?
I am after that logic.
Pegasystems Inc.
US
Hello!
Could you please share your SR ID so that we can connect it with this thread?
Thanks!
Marissa | Senior Moderator | Pega Collaboration Center