Currently the output of the StatsCollector is only stored in the internal database and processed by the usage server (if configured).

Optionally output all the Stats collected by the statscollector to Graphite and have the admin process the data there.

A global configuration option allows it to be configured: stats.output.uri: graphite://1.2.3.4:2003/prefix

VmStats will be formatted this way:

  • <prefix>.cloudstack.stats.instances.<vm uuid>.cpu.num
  • <prefix>.cloudstack.stats.instances.<vm uuid>.cpu.utilization
  • <prefix>.cloudstack.stats.instances.<vm uuid>.network.read_kbs
  • <prefix>.cloudstack.stats.instances.<vm uuid>.network.write_kbs
  • <prefix>.cloudstack.stats.instances.<vm uuid>.disk.read_kbs
  • <prefix>.cloudstack.stats.instances.<vm uuid>.disk.write_kbs
  • <prefix>.cloudstack.stats.instances.<vm uuid>.disk.read_iops
  • <prefix>.cloudstack.stats.instances.<vm uuid>.disk.write_iops

This has been implemented with commit: https://git-wip-us.apache.org/repos/asf?p=cloudstack.git;a=commitdiff;h=840c2fda852b86f51245a31e7cd91cdb0ad4bf49

  • No labels