Process Metrics

The following metrics are aggregated and reported for each process or process class (except Total Process Count, which measures all processes observed by the agent). See Machine Agent Settings for Server Visibility to modify the default process monitoring.


Metric Name	Description	Basic or Server Visibility	Default Monitoring Mode
Count	The number of processes in this class consuming CPU or memory resources or the total number of processes in this class.	Server Visibility	KPI
CPU Used (%)	Percentage of the CPU bandwidth used by all processes in a process class. A process using 100% CPU is executing on all processors on the system. If there are four cores on the machine and the process is executing four threads, each executing on one core, then the process can use up to 400% CPU. If there are four cores on the machine and the process is executing one thread on one core, then the process can use up to 25% CPU.	Server Visibility	KPI
Major Faults/sec	Number of major page faults caused by all processes in this class.	Server Visibility	Diagnostic
Minor Faults/sec	Number of minor page faults caused by all processes in this class.	Server Visibility	Diagnostic
Memory Used (%)	Percentage of memory consumed by the top 10 consuming processes or the percentage of memory used by all processes in this class.	Server Visibility	KPI
Memory Used (KB)	Amount of memory used by all processes in this class.	Server Visibility	Advanced
Memory Virtual (KB)	Current size of the virtual address space that the process is using. Use of virtual address space does not necessarily imply a corresponding use of either disk or main memory pages. Virtual space is finite and, by using too much, the process can limit its ability to load libraries.	Server Visibility	Advanced
Thread Count	The number of kernel threads used by all processes in this class.	Server Visibility	Diagnostic
Total Process Count	All individual processes observed by the Machine Agent,beforeit filters out unreported processes and groups the remaining processes into classes. You can use this metric to identify anomalies in the number of processes running on a specific machine. For example, if the average Total Process Count on a machine is usually 200 to 300, and then suddenly rises to 2000 or more, this could indicate a problem on that machine. To collect this metric, the Machine Agent captures the number of processes every 30 seconds and calculates the average number of processes per minute. Observed is the average number of processes per minute. Min and Max are the 30-second buckets with the least and most processes. Count and Sum are the number of process captures each minute and the sum of the processes in each capture for that minute (capture1 + capture2).	Server Visibility	KPI

¹See Linux Kernel Processes.

²When collecting Process metrics on Solaris, the Machine Agent observes and captures only the first 80 characters of each process name and argument list. The agent considers only the first 80 characters of each process string when it applies allowlists ("always monitor this process") and blocklists ("never monitor this process") defined in ServerMonitoring.yml.

³ See Default Monitoring Mode.