Memory monitoring on AIX using SystemEDGE

Document ID : KB000077072
Last Modified Date : 26/04/2018
Show Technical Document Details
Question:
We have enabled memory monitoring on AIX servers via the physical memory utilization variable from eHealth live exceptions, but we are receiving complaints that those are false positive alerts.

We would like to know what is the ideal way to enable memory alarms on AIX servers using SystemEDGE?
Answer:
  • The way in which AIX manages memory is different than on Linux.
  • AIX will generally use most of the total memory default.
  • On AIX SystemEDGE cannot include buffers and cached memory in the free memory calculation like on Linux ( linux_freemem_include buffers cached)
  • You cannot use memory utilization (%) to effectively monitor memory utilization on AIX like with Linux.
  • Monitoring Swap metrics is a more appropriate gauge for AIX:
• swapCapacity -> kernelperf.30 (1.3.6.1.4.1.546.1.1.7.8.30) 
Type: INTEGER Access: read-only 
The percentage of the system's total swap in use. The value of this object can be computed by dividing swapInUse(13) by totalSwapSpace(29) and multiplying by 100. It is included here for convenient polling, monitoring, and history sampling. This object is an aggregate value over all the swap partitions and areas. 

• numSwapIns -> kernelperf.20 (1.3.6.1.4.1.546.1.1.7.8.20) 
Type: Counter Access: read-only 
The total number of processes that have been swapped in since the kernel was last initialized. Processes are swapped to and from secondary storage when memory is in short supply. A high rate may indicate an overloaded system. 

• numSwapOuts -> kernelperf.21 (1.3.6.1.4.1.546.1.1.7.8.21) 
Type: Counter Access: read-only 
The total number of processes that have been swapped out since the kernel was last initialized. Processes are swapped to and from secondary storage when memory is in short supply. A high rate may indicate an overloaded system. 
  • You can also leverage SystemEDGE MIB extension functionality if you find a native command that returns the correct information for your needs. 
  • You can roll the command into a script and have the agent send notifications based on the script's return.
  • An example of these commands would be:
"vmstat -v |grep -e "numperm percentage" -e "percentage of memory used for computational pages"

24.6 numperm percentage
67.7 percentage of memory used for computational pages


 
Additional Information: