*** openstackstatus has joined #openstack-telemetry | 00:44 | |
*** ChanServ sets mode: +v openstackstatus | 00:44 | |
*** psachin has joined #openstack-telemetry | 03:37 | |
*** tesseract has joined #openstack-telemetry | 08:08 | |
*** bengates has joined #openstack-telemetry | 08:09 | |
*** witek_ has joined #openstack-telemetry | 09:13 | |
*** saska has joined #openstack-telemetry | 09:14 | |
saska | Hi, does anybody know whether is it possible for Monasca to collect metrics mentioned in the following link? https://opendev.org/openstack/watcher/src/branch/master/watcher/decision_engine/datasources/monasca.py#L31 If it is not possible, what is missing in Monasca's code (or architecture) which would make it possible? | 09:20 |
---|---|---|
witek_ | saska: some of these metrics are available via native Monasca checks | 09:25 |
witek_ | https://github.com/openstack/monasca-agent/blob/master/docs/Plugins.md#agent-plugin-detail | 09:26 |
witek_ | https://github.com/openstack/monasca-agent/blob/master/docs/Libvirt.md#per-instance-metrics | 09:26 |
witek_ | alternatively you can collect any Ceilometer or Prometheus measurements | 09:27 |
saska | witek_ : thank you. In the "per instance metrics" section I noticed 2 which may be helpful. mem.free_perc which is described as " Percent of memory free" and mem.resident_gb which is described as "Total memory used on host, an Operations-only metric ". Does the first one refer to 1 VM (per instance== 1 VM)? What does it mean "operations-only metrics"? | 09:51 |
witek_ | saska: yes, instance = VM, this document describes the libvirt plugin | 09:59 |
witek_ | this plugin provides two sets of measurements, one for the operator (operator project) and one for the tenant | 10:00 |
witek_ | mem.resident_mb is available only in operator project | 10:01 |
witek_ | please also see this section https://github.com/openstack/monasca-agent/blob/master/docs/Libvirt.md#untunable-metrics | 10:02 |
saska | witek_ : thank you. The last link shows that for an administrator, metrics are collected per instance (in libvirt plugin) and for a user per tenant. And the vm.mem.resident_mb is available only for open stack operator since a user (owner of a tenant) does not need to know how many gb of memory does his project occupy on a host. Is that right? | 10:38 |
witek_ | exactly | 10:38 |
*** witek_ is now known as witek | 10:38 | |
saska | witek_ : thank you. I appreciate your help. | 10:51 |
witek | saska: you're welcome, don't hesitate to ask more | 10:54 |
saska | witek: Thank you. I have many, but trying to figure out some of them on my own. What is the difference between vm.mem.total_mb and vm.mem.resident_mb? Is it possible, that a VM has allocated memory on more than 1 host? | 11:25 |
witek | here is libvirt documentation describing these values: https://www.libvirt.org/html/libvirt-libvirt-domain.html#virDomainMemoryStatTags | 11:34 |
witek | mem.total_gb - AVAILABLE | 11:34 |
witek | mem.resident_db - RSS | 11:35 |
*** saska has quit IRC | 11:54 | |
*** rcernin has quit IRC | 12:25 | |
*** tkajinam has quit IRC | 12:27 | |
*** psachin has quit IRC | 12:47 | |
*** psachin has joined #openstack-telemetry | 13:12 | |
*** witek has quit IRC | 13:47 | |
*** bengates has quit IRC | 13:57 | |
*** witek has joined #openstack-telemetry | 14:04 | |
*** bengates has joined #openstack-telemetry | 14:12 | |
*** psachin has quit IRC | 14:52 | |
*** tesseract has quit IRC | 16:40 | |
*** bengates has quit IRC | 16:50 | |
*** bengates has joined #openstack-telemetry | 16:51 | |
*** witek has quit IRC | 16:52 | |
*** bengates has quit IRC | 16:56 | |
*** witek has joined #openstack-telemetry | 17:05 | |
*** witek has quit IRC | 17:15 | |
*** rcernin has joined #openstack-telemetry | 22:04 | |
*** rcernin has quit IRC | 22:04 | |
*** rcernin has joined #openstack-telemetry | 22:05 | |
*** etp has quit IRC | 22:15 | |
*** etp has joined #openstack-telemetry | 22:16 | |
*** tkajinam has joined #openstack-telemetry | 22:56 | |
*** dave-mccowan has joined #openstack-telemetry | 23:10 | |
*** threestrands has joined #openstack-telemetry | 23:50 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!