INC-175205 · Issue 652529
Agent restart made more resilient
Resolved in Pega Version 8.4.5
During a timeout outage, agents intermittently encountered a NoClassDefFound error and were disabled. After the database connection was restored, the agents were not automatically restarting. This has been resolved by making the code more resilient to errors; agent will be restarted when a temporary issue is detected, and the restart will be immune to most exceptions.
INC-179861 · Issue 658948
AES call to mbean disabled
Resolved in Pega Version 8.4.5
Under certain conditions, the AESRemote 'push memory' agent call to the legacy memory management mbean caused unexpected behavior in declarative rules for work running in data flows and queue processors. This has been resolved at this time by overriding the push memory usage agent activity to skip mbean and connect REST.
INC-180207 · Issue 659097
Corrected race condition that caused JobScheduler error
Resolved in Pega Version 8.4.5
A race condition between the pzInitializeQueueProcessors and pyPersistNodeState job schedulers that both run at platform's start was causing a null pointer exception with the error "com.pega.platform.executor.jobscheduler.scheduler.JobExecutionException: Activity execution failed with [PRRuntimeException]". This has been resolved by moving the QueueProcessor initialization process from JobScheduler to StartupTask.