The OpenStack Monitoring Plugin gathers performance statistics provided by the Ceilometer and Nova services from the OpenStack virtual cloud VMs. Gathered measures are configurable and can be setup dynamically depending on VM flavors.
Name and Version
OpenStack Monitoring Plugin 0.9.8.1
Eugene Turetsky (firstname.lastname@example.org)
The OpenStack Monitoring Plugin gathers performance statistics from the OpenStack virtual cloud VMs
2015-01-30 Initial Release
Import the plugin jar files into the dynaTrace Server. For details how to do this please refer to the dynaTrace documentation.
For details about plugin see this document.
The plugin allows to gather performance statistics from the OpenStack virtual cloud VM(s). It provides statistics from the following metric groups: Status, CPU, Memory, Disk I/O, Network I/O, Swap, and HostMetrics.
The Status metric group contains Server Status metric and Server Task State metric. The Server Status metric could have one of the following values (see vm_state here):
ACTIVE=0, BUILD=1, ERROR=2, RESIZE=3, SHUTOFF=4, UNKNOWN=5, MIGRATING=6, INITIALIZED=7, RESCUED=8, PAUSED=9, SUSPENDED=10, STOPPED=11, SOFT_DELETED=12, HARD_DELETED=13, RESIZED=14, REBUILDING=15, RESIZING=16.
The Server Task State metric could have one of the following values (see task_state here):
NONE=0, BUILDING=1, IMAGE_SNAPSHOTTING=2, IMAGE_BACKINGUP=3, UPDATING_PASSWORD=4, PAUSING=5, UNPAUSING=6,
SUSPENDING=7, RESUMING=8, DELETING=9, STOPPING=10, STARTING=11, RESCUING=12, UNRESCUING=13, REBOOTING=14,
REBUILDING=15, POWERING_ON=16, POWERING_OFF=17, RESIZING=18, RESIZE_REVERTING=19, RESIZE_CONFIRMING=20,
SCHEDULING=21, BLOCK_DEVICE_MAPPING=22, NETWORKING=23, SPAWNING=24, RESIZE_PREP=25, RESIZE_MIGRATING=26,
RESIZE_MIGRATED=27, RESIZE_FINISH=28, RESIZE_VERIFY=29, UNKNOWN=30
The HostMetrics group provide CPU, Disk, and Memory allocation statistics for the server which hosts virtual cloud VMs. Other metric groups provide statistics from the Ceilometer and the Nova services.
Names of measures which are gathered by the plugin are composite names. Naming convention is: <server-name>"|"<nova-measure-name> or <ceilometer-measure-name>"|"<tenant-name>.
Note that if there are many VMs on the virtual cloud the number of gathered by the plugin measures will be significant. Hence this could potentially cause measure explosion. Please configure Dynatrace Performance Warehouse appropriately and choose only necessary for the analysis measures.
The following screenshot below depicts plugin's configuration parameters They are:
- Tenant (project);
- User - needs to have "admin" role;
- Host Name - name of the server which hosts virtual cloud;
- Authentication URL (Keystone service);
- Nova URL (Nova service);
- Ceilometer URL (Ceilometer service);
- Mapping Measures File - file which contains gathered measures and their mappings to the base measures and metric groups. See here example of this file;
- Web API Version - default v2;
- Test Mode - puts measures which are not present in the Mapping Measures File to the "OtherMeasuresGroup".
Default List of Metric Groups supported by the plugin
The following screenshot depicts default list of metric groups that plugin can support:
- Users are welcome to add or remove metric groups depicted in this section.
Ceilometer virtual cloud VM measures
The following screenshot depicts example of metrics that plugin is capturing from the Ceilometer service:
Please find examples of dashlets with Ceilometer measures in the Examples section below.
Nova virtual cloud VM measures
The following screenshot depicts example of metrics that plugin is capturing from the Nova service:
- Except memory and host usage measures, measures from the Nova service are cumulative. The rate attribute in the measure tag allows to get derived rate measure from the given measure tag. See examples of the rate attribute in the above screenshot for errors and Swap In/Out measures.
Please find examples of dashlets with Nova measures in the Examples section below.
Host Usage measures
The host statistics are computed by using information about the flavors of the instances that run on the host identified by the Host Name parameter. These statistics include:
- The CPU measure shows the sum of the virtual CPUs for instances running on the host;
- The Memory measure shows the sum of the memory (in MB) allocated to the instances that run on the host;
- The Disk measure shows the sum of the root and ephemeral disk sizes (in GB) of the instances that run on the host.
The dynamic measure
used_now shows the sum of the resources allocated to the instances that run on the host, plus the resources allocated to the virtual machine of the host itself. The dynamic measure
used_max shows the sum of the resources allocated to the instances that run on the host. For more information please follow this link: summary of resource usage of all of the instances running on the host.
Next screenshot depicts Host CPU statistics:
Next screenshot depicts Host Memory statistics:
Next screenshot depicts Host Disk statistics:
Examples of Dashlets
Below screenshot shows an example of servers availability dashlet. There are 4 servers for the "demo" tenant in this example: nano, micro, tiny, and small. The first two servers, nano and micro, became ACTIVE between 7:54am and 8:06am timeframe while the last two, tiny and small, are in the SHUTOFF state.
Next screenshot shows Disk I/O Nova usage statistics gathered by the plugin:
Next screenshot shows Disk I/O Ceilometer usage statistics gathered by the plugin:
Next screenshot shows example of the Network I/O Nova statistics gathered by the plugin:
Next screenshot shows Nova memory usage statistics gathered by the plugin:
Next screenshot shows CPU cumulative Ceilometer and Nova statistics gathered by the plugin:
Next screenshot shows CPU usage Ceilometer statistics gathered by the plugin: