Performance alert U00005900 Forced trace count detected Automation Engine.

Products

• Automic Automation

Issue/Introduction

The Automation Engine logs the message U00005900 Performance issue 'Forced trace count' was detected. This article explains why this message appears and how to identify the root cause.

Symptoms

• The message U00005900 Performance issue 'Forced trace count' was detected appears in the Java Work Process (JWP) logs.
• The log summary indicates a ratio of 100% and a threshold of 0.0%.

Environment

• Automation Engine (all versions)
• Java Work Process (JWP)

Cause

The Automation Engine's internal Performance Monitoring (PM) subsystem flags this issue when an unhandled exception occurs. When the JWP logs an unhandled exception via LogComponent.logError(), the system forces a diagnostic trace dump, even if tracing is normally disabled. The PM subsystem monitors the frequency of these forced traces and alerts when the count exceeds the configured threshold.

Resolution

1. Search the JWP logs for message 45014 or nearby error/exception stack traces that correlate with the timestamp of the U00005900 alert.
2. Locate the forced trace dump files in the system's trace directory generated at those timestamps.
3. Analyze the stack dump within the trace files to identify the component, job, or connection causing the unhandled exception.
4. Address the underlying error identified in the stack trace.
5. If further adjustment is required, add an explicit FORCED_TRACE_COUNT entry to the UC_PERFORMANCE_THRESHOLDS VARA object in client 0 to adjust the alert sensitivity.
6. Restart the non Java-based CP processes.

Note: If the issue persists after troubleshooting the underlying exceptions, Contact Broadcom Support for further assistance.