@PradeepPydi  

Did you estimate the hardware size for your Pega applications on Pega 8.6.3?  Has this recently changed?

Please go through below url  https://community.pega.com/knowledgebase/documents/hardware-sizing-estimate-request

PegaJdbcURLHandler internally creates an instance of VersionedJdbcJarReader and registers with MBean Server. While registering, it appends the hashcode of PegaJdbcURLHandler instance to bean name. Internally every reader maintains an in-memory cache of engine classes information. 

If this needs a thorough investigation of the  memory utilization for the server, then we would require you to log a support incident for this.  

Your logs may show requestor lock exceptions ,thread dumps etc so the engineers would need to identify some of the bottlenecks.

What do your  alert logs show?  If they  contains   PRGA0086 and PEGA0019 alerts  you would need to provide : 

1)   jvm arguments . 2)  GC logs.  If you haven't enabled GC could you  add below jvm arguments

-verbose:gc -Xloggc:"C:\tomcat722a\logs\tomcat722a_gc.txt" -XX:HeapDumpPath="C:\tomcat722a\logs\heapdump.hprof" -XX:+HeapDumpOnOutOfMemoryError -XX:+PrintAdaptiveSizePolicy -XX:+PrintGCDetails -XX:+PrintCompilation -XX:+PrintGCTimeStamps -XX:+PrintGCDateStamps -XX:+PrintGCApplicationConcurrentTime -XX:+PrintGCApplicationStoppedTime -XX:+PrintTenuringDistribution -XX:+PrintCommandLineFlags

NOTE - some of these parameters may be incompatible with older JVM versions; therefore, validate these parameters both through documentation for your specific JVM variant, and through lab testing.

- depending upon version/compatibility, some of these parameters may cause a JVM to fail startup.

If you want to rotate the GC logs , you can add below jvm argument

-XX:+UseGCLogFileRotation -XX:NumberOfGCLogFiles=3 -XX:GCLogFileSize=2M

Enabling GC in production is highly recommended by IBM JVM experts to troubleshoot the performance ,there is little overhead for VerboseGC .   Please have a look at below article

https://www.ibm.com/developerworks/community/blogs/troubleshootingjava/entry/verbose_gc_performance?lang=en

Note :Please test these arguments in one of your development environment followed by production .

Could you please take the thread dumps if you are seeing the slowness:

 Monitor the system and capture the following when issue reoccurs.     a. a set of thread dumps     b. enable debug level logging on following loggers on same node where thread dumps are being collected. reset the loggers once the thread dumps collection finished

        com.pega.pegarules.data.internal.dictionary.ClassMapImpl         com.pega.pegarules.data.internal.dictionary.ClassMapImpl.ClassToTableMap         ClassMapChangeLog

If you decide to log a support incident, they will require: 1)Pega Rules 2)Pega Alerts 3)GC logs 4)Systemerr 4)Systemout 5) Thread dumps 6) infrastructure details in terms of RAM, CPU cores ..    .

Please let us know  the support ticket ID if you decide to log one.