Monit memory utilization different from Ubuntu's own diagnostics

Question

I have a Linode 512 VPS running Ubuntu 12.04 and I noticed a very big discrepency between Monit's report of memory utilization and Ubuntu's report. This becomes an issue when monit tries to restart some of my processes when it crosses a certain threshold. Does anyone know why there is this discrepency?

Answer 1

Without digging too deep into monit's source code, it appears that monit takes into account memory buffers used by the kernel when computing used memory.

From monit on my Linode:

  status                            Running
  monitoring status                 Monitored
  load average                      [0.00] [0.01] [0.05]
  cpu                               0.0%us 0.0%sy 0.0%wa
  memory usage                      264652 kB [52.0%]
  swap usage                        47096 kB [2.2%]
  data collected                    Tue, 11 Dec 2012 20:51:45

and running free right after:

               total       used       free     shared    buffers     cached
  Mem:        508576     405572     103004          0      20288     120636
  -/+ buffers/cache:     264648     243928
  Swap:      2097148      47092    2050056

Considering monit's default configuration is to refresh every 120 seconds, 264648 kB pretty accurate compared to monit's 264652 kB.