This article describes the correct way to interpret memory usage statistics that are reported on UNIX based systems. This is particularly useful when setting up monitoring mechanisms to track memory usage on Appian machines and setting upper limits to trigger alerts when the usage crosses this threshold.
This concept is best understood with an example. Consider the following output when running the standard command free -m to check memory usage:
$ free -m total used free shared buffers cached Mem: 1504 1491 13 0 91 764 -/+ buffers/cache: 635 869 Swap: 2047 6 2041
At a first glance on the total and used columns, it would seem that the memory usage on the server is at 99% of total memory. However note that in addition to standard free memory, there is also free cached memory which also needs to be taken into account to get an accurate picture of total free memory on the server calculated as follows:
Total free memory = free + cache
In the above example, the total free memory would be 882M which comes to over 58% of available memory.
More information on the various types of memory in linux can be found in this site: http://www.linuxatemyram.com/
This article applies to all versions of Appian.
Last Reviewed: February 2017