Under normal conditions with the Global Manager process memory size well within the operating system limit, monitor memory size to establish a baseline that can be used to detect performance issues early.
For UNIX systems, this can be automated by creating a script that periodically executes ps -l -p proclist and writes the output to a file.
Investigating the values from the file helps to understand the memory growth pattern over time. Once the pattern is understood, it can be used to establish a baseline value for the memory usage of a running Global Manager.
The baseline value is then used to establish a memory usage threshold. Using a monitoring script, an alert can be generated if the memory used by the Global Manager ever exceeds the threshold.
Once the alert is received, investigate the queue growth before attempting to remedy the condition.