vhostmd.xml

<?xml version="1.0" ?>
<!DOCTYPE vhostmd SYSTEM "vhostmd.dtd">

<!--

Copyright (C) 2008 Novell, Inc.

Configuration file for virtual host metrics daemon (vhostmd).

A metrics disk between 1024 bytes and 256Mbytes is supported.

Supported metric types are: int32, uint32, int64, uint64, real32,
real64, and string

A metric's value is set to the output produced by executing its action.

'action' can include the special token NAME, in which case the name of
the vm currently under inspection is substituted for NAME.  Only useful
within the vm element.

NOTE - 'action' must be a valid shell builtin, script or external
command found in the path specified by the global <path> element.
When chaining commands, '&', '<' and '>' are reserved characters,
therefore '&amp;', '&lt;' and '&gt;' must be used instead. For example,
the logical && operator must be replaced with "&amp;&amp;".

-->

  <vhostmd>
    <globals>
      <disk>
        <name>host-metrics-disk</name>
        <path>/dev/shm/vhostmd0</path>
        <size unit="k">256</size>
      </disk>
      <virtio>
        <max_channels>1024</max_channels>
        <expiration_time>15</expiration_time>
      </virtio>
      <update_period>5</update_period>
      <path>/usr/sbin:/sbin:/usr/bin:/bin:/usr/share/vhostmd/scripts</path>
      <transport>vbd</transport>
      <transport>virtio</transport>
      <!-- <transport>xenstore</transport> -->
    </globals>
    <metrics>
      <metric type="string" context="host">
        <name>HostName</name>
        <action>
          virsh CONNECT hostname | tr -d '[:space:]'
        </action>
      </metric>
      <metric type="string" context="host">
        <name>VirtualizationVendor</name>
        <action>
	  rpm -q --qf '%{VENDOR}\n' -qf /etc/os-release
        </action>
      </metric>
      <metric type="string" context="host">
        <name>VirtualizationProductInfo</name>
        <action>
	  virsh version | awk '/Running hypervisor/ {print $(NF-1),$NF}'
        </action>
      </metric>
      <metric type="uint32" context="host">
        <name>TotalPhyCPUs</name>
        <action>
          virsh nodeinfo | awk '/^CPU\(s\)/ {print $2}'
        </action>
      </metric>
      <metric type="uint32" context="host">
        <name>NumCPUs</name>
        <action>
          virsh nodeinfo | awk '/^CPU\(s\)/ {print $2}'
        </action>
      </metric>
      <metric type="uint64" context="host" unit="MiB">
        <name>TotalPhyMem</name>
        <action>
          echo $((`virsh nodeinfo | awk '/^Memory/ {print $3}'` / 1024))
        </action>
      </metric>
      <metric type="uint64" context="host">
        <name>UsedMem</name>
        <action>
          [ -f /proc/xen/privcmd ] &amp;&amp; echo "$((`xentop -b -i 1 | awk '/Domain-0/ {print $5}'` / 1024))" || \
          free | egrep -i '^[[:space:]]*(.*buffers/cache:)' | awk '{ printf "%d\n", $3/1024; }'
      </action>
      </metric>
      <metric type="uint64" context="host">
        <name>FreeMem</name>
        <action>
          [ -f /proc/xen/privcmd ] &amp;&amp; xl info | awk '/^free_memory/ {print $3}' || \
          free | egrep -i '^[[:space:]]*(.*buffers/cache:)' | awk '{ printf "%d\n", $4/1024; }'
        </action>
      </metric>
      <metric type="uint64" context="host">
        <name>PagedInMemory</name>
        <action>
          vmstat -s | awk '/pages paged in/ {printf "%d\n", $1 / 1024 * $(getconf PAGESIZE) / 1024;}'
        </action>
      </metric>
      <metric type="uint64" context="host">
        <name>PagedOutMemory</name>
        <action>
          vmstat -s | awk '/pages paged out/ {printf "%d\n", $1 / 1024 * $(getconf PAGESIZE) / 1024;}'
        </action>
      </metric>
      <metric type="group" context="host">
        <name>PageRates</name>
        <action>pagerate.pl</action>
        <variable name="PageInRate" type="uint64"/>
        <variable name="PageFaultRate" type="uint64"/>
      </metric>
      <metric type="real64" context="host" unit="s">
        <name>TotalCPUTime</name>
        <action>
          [ -f /proc/xen/privcmd ] &amp;&amp; xl list | awk '/^Domain-0/ {print $6}' || \
          awk '
          function user_hz(   hz)
          {
            cmd = "getconf CLK_TCK";
            cmd | getline;
            hz  = $1;
            close(cmd);

            return hz;
          }

          BEGIN {
            USER_HZ      = user_hz();
            TotalCPUTime = 0;

            while ( 0 &lt; ( getline &lt; "/proc/stat" ) )
            {
              if ( "cpu" == $1 )
              {
                TotalCPUTime = $2 + $3 + $4;

                break;
              }
            }
            close("/proc/stat");

            TotalCPUTime /= USER_HZ;
            printf "%f\n", TotalCPUTime;
          }'
        </action>
      </metric>
      <metric type="real64" context="vm" unit="s">
        <name>TotalCPUTime</name>
        <action>
          virsh CONNECT dominfo NAME | sed 's/: */:/' | \
          awk -F: '/CPU time/ {print $2;}' | sed 's/s//'
        </action>
      </metric>
    </metrics>
  </vhostmd>