Flexiant Data Collector

Provided Metrics

Resources information and relationships

The information about resources in the Flexiant Testbed (Nodes, Racks and Clusters) are contained in configuration files packaged together the data collector and available here. Such files also contain information about their relationships according to the Tower 4Clouds meta-model. Custom files can be provided as described in this documentation.

The following conventions were used in the development of this data collectors when the model is created:

• Nodes are given their own IP as id, with dots replaced by underscores. Type is provided by the configuration.
• The updated list of VMs contained in a Node is retrieved via the Flexiant provided API. The important assumption here is that VMs ids in the Tower 4Clouds model are equal to their IP address, with dots replaced by underscores. If such assumption is not valid, higher level metrics such as Response Time, collected by other data collectors, cannot be grouped by resources monitored by this data collector (e.g. by Cluster).
• Clusters and Racks are given the id provided by the Flexiant provided API.

Usage

Flexiant DC can be configured by means of different options (latters replaces the formers):

• Default configuration
• Environment Variables
• CLI arguments

Environment Variables

If these variables aren't setted the DC will use the default values specified in the table.

CLI Arguments

CLI Arguments overwrite default value or environment variable for any parameter.

Usage: java -jar DC.jar [options]
  Options:
    -managerip
      Manager IP address
    -managerport
      Manager port
    -config-file
      Config file path
    -nodes-file
      Nodes file path
    -clusters-file
      Clusters file path
    -racks-file
      Racks file path
    -placement-file
      Placement file path

Configuration File (config-file)

The config file contains a lot of properties that specify URLs of some remote files.
Below you find the structure of the file with currently used url:

#Flexiant DC Properties
URL_CPU_METRIC=https://cp.sd1.flexiant.net/nodecpu10/
URL_RAM_METRIC=https://cp.sd1.flexiant.net/noderam10/
URL_NODELOAD_METRIC=https://cp.sd1.flexiant.net/nodeload10/
URL_TXNETWORK_METRIC=https://cp.sd1.flexiant.net/nodenet10/
URL_RXNETWORK_METRIC=https://cp.sd1.flexiant.net/nodenet10/
URL_STORAGE_METRIC=https://cp.sd1.flexiant.net/storage10/
URL_RACKLOAD_METRIC=https://cp.sd1.flexiant.net/rackload10/upsload.csv
URL_VMS=https://cp.sd1.flexiant.net/VMPlacement/FCOVMPlacement.csv

The properties specify URL of files which contain metrics' samples.

Nodes, Clusters and Racks file

You can specify the resources which DC have to monitor by create three csv file (nodes, clusters and racks). Once files are created you have to pass the path of files using one of the ways exaplined above. Below there is an example of the structure of the files:

Id,Type
nodeId,NodeType

Id,Type
rackId,RackType

Id,Type
clusterId,ClusterType

Rack/node, Cluster/node relations file (placement-file)

If you want you can specify the relationships between racks and nodes and between clusters and nodes, you can create a .CSV file which contains that information. Below there is an example of the structure of the file:

Rack,Node,Cluster
A5,10.158.128.15,Cluster2
A5,10.157.128.31,Cluster1
A5,10.158.128.11,Cluster2
A5,10.158.128.16,Cluster2

An example of rule

Here we provide an example of rule that can be used to monitor the average CPU utilization of all nodes in the Flexiant clusters grouped by Cluster.

<monitoringRules xmlns="http://www.modaclouds.eu/xsd/1.0/monitoring_rules_schema">
    <monitoringRule id="CpuUtilizationRule" startEnabled="true" timeStep="5" timeWindow="5">
        <monitoredTargets>
            <monitoredTarget class="Node" />
        </monitoredTargets>
        <collectedMetric metricName="CPUUtilization">
            <parameter name="samplingTime">10</parameter>
        </collectedMetric>
        <metricAggregation aggregateFunction="Average" groupingClass="Cluster"/>
        <actions>
            <action name="OutputMetric">
                <parameter name="resourceId">ID</parameter>
                <parameter name="metric">AverageCPUUtilization</parameter>
                <parameter name="value">METRIC</parameter>
            </action>
        </actions>
    </monitoringRule>
</monitoringRules>