High RAM usage due to job over-prediction in AAI

Products

• Automation Analytics & Intelligence

Issue/Introduction

Users may experience the following issues when running Automation Analytics & Intelligence (AAI) version 24.4.0:

Symptoms

• High RAM utilization (90%+) shortly after installation or upgrade.
• Server CPU saturation (96–100%) occurring in recurring bursts (e.g., every 4–10 minutes).
• Users are unexpectedly redirected to the login screen while active.
• jaws.log contains QueryTimingService warnings with values above 100% and high predicted run counts (e.g., > 1 million runs).
• Slow performance when processing file trigger or file watcher job chains.

Environment

• Product: Automation Analytics & Intelligence (AAI)
• Version: 24.4.0
• Scheduler: AutoSys (with complex file watcher dependency chains)

Cause

A product defect (DE194647) causes the JammerSequence prediction engine to over-predict simulated job runs for certain dependency configurations.

Resolution

his issue is resolved in version 24.4.1 and higher.

IMPORTANT:
Fixed high RAM usage due to over prediction. The stats.useTerminateJobRunsInStatistics server parameter must be set to true (default = false) and a Statistics Generator run must occur for the improvements to take effect.