{section}
{column}

Two kinds of graph can be generated by Cacti:
* using a simple data: data are returned by the probe and are directly and completely graphed,
* or using a data query: data are returned by the probe and must be filtered before to be graphed

{note}The following steps by steps process must be adapted to your metric{note}

{column}
{column:width=350px}
{panel:title=Table of contents}{toc:outline=true}{panel}
{panel:title=Contributors}{contributors:order=name|mode=list|showAnonymous=true|showCount=true|showLastTime=true}{panel}
{column}
{section}

{anchor:Creating a graph using a simple data}
h1. Using a simple data

This is done using a simple *script/command* based on Petals CLI to collect metric values.

Required input parameters are:
* the command line to use with its arguments and input fields,
* output fields

To create such a graph, please follow the following steps:

!petals-container-localtransporter-dim-petals-host.jpg|align=center!
As the Petals node hostname can be automatically set from the hostname of the monitored host, the field *Special Type Code* is set.
# Add others *input fields*
!petals-container-localtransporter-dim-petals-jmx-port.jpg|align=center!
As for the JMX port, create an input field for the Petals JMX username and the Petals JMX user password.
# Add the *output fields* returned by the script
For each output fields of the script/command (here, in the sample, for each message exchange status "Active", "Done" and "Error"), create an output field:
!petals-container-localtransporter-dim-output-active.jpg|align=center!
# Check that you have the following *Data Input Method*:
!petals-container-localtransporter-dim-full.jpg|align=center!
# Create a new *Data Template*:
!petals-container-localtransporter-dt.jpg|align=center!
Select your data input method previously created as *Data Input Method*.
Add your first *data item* associated to an output field of the script/command previously created (here in the sample, the number of delivered messages for message exchange status 'Active'). The *data source type* of the item must be correctly set according to the output field type.
# Create the others *data items* associated to the remaining output fields
!petals-container-localtransporter-dt-items.jpg|align=center!
# A new block *Custom data* appears, check *Use Per-Data Source Value* for all except for the Petals node hostname
!petals-container-localtransporter-dt-custom.jpg|align=center!
# Create a new *Graph Template*:
!petals-container-localtransporter-gt.jpg|align=center!
# Add a *Graph Item* to graphed one of the previous *data item* (here in the sample, the number of delivered messages for message exchange status 'Active')
!petals-container-localtransporter-gt-active.jpg|align=center!
# Add a *Graph Item* to displayed the legend of your *data item* (here in the sample, the number of delivered messages for message exchange status 'Active')
!petals-container-localtransporter-gt-active-legend.jpg|align=center!
# Add others *Graph Items* associated to the remaining *data items*
!petals-container-localtransporter-gt-full.jpg|align=center!
# Create a *New Graph*
!petals-container-localtransporter-ng.jpg|align=center!
# Set parameters needed to the *Data Input Method*
!petals-container-localtransporter-ng-params.jpg|align=center!

The new graph can be viewed after several minutes needed to collect data. By default a collect is done every 5 minutes. To view the graph, go into *Graph management"

{anchor:Creating a graph using a data query}
h1. Using a data query

!petals-delivered-messages-by-endpoint.png|align=center!

This is done using a *data query* to collect metric values.

Required input parameters are:
* the data query,
* the output fields.

The data query definition of our sample to put in the XML file is:
{code}
<interface>
<name>Get Delivered Messages by the Local transporter to an endpoint</name>
<script_path>/home/cdeneux/petals-cli-2.1.1-SNAPSHOT/bin/petals-cli.sh</script_path>
<arg_prepend>-h localhost -n 7700 -u petals -p petals -c -- monitoring -o local-transporter -f delivered-messages</arg_prepend>
<arg_index>-- --list-endpoints</arg_index>
<arg_query>-- --query-endpoints</arg_query>
<arg_get>-- --query-endpoints</arg_get>
<output_delimeter>!</output_delimeter>
<index_order>endpointName</index_order>
<index_order_type>alphanumeric</index_order_type>
<!--index_title_format>|chosen_order_field|</index_title_format-->

<fields>
<endpointName>
<name>Endpoint name</name>
<direction>input</direction>
<query_name>NAME</query_name>
</endpointName>
<activeDeliveredMsgs>
<name>Active Message exchange delivered counter</name>
<direction>output</direction>
<query_name>ACTIVE</query_name>
</activeDeliveredMsgs>
<doneDeliveredMsgs>
<name>Done Message exchange delivered counter</name>
<direction>output</direction>
<query_name>DONE</query_name>
</doneDeliveredMsgs>
<errorDeliveredMsgs>
<name>Error Message exchange delivered counter</name>
<direction>output</direction>
<query_name>ERROR</query_name>
</errorDeliveredMsgs>
</fields>

# First, create a new *Data Query*:
!petals-container-localtransporter-itf-dq.jpg|align=center!
# Go to the *device* for which you want collect metric values, and associate it to the data query previously created
!petals-container-localtransporter-itf-device-associate.jpg|align=center!
# Create a new *Data Template*:
!petals-container-localtransporter-itf-dt.jpg|align=center!
Select "*Get Script Data (Indexed)*" as *Data Input Method*.
Add your first *data item* associated to an output of the data query (here in the sample, the number of delivered messages for message exchange status 'Active'). The *data source type* of the item must be correctly set according to the output field type.
# Create others *data items* associated to the remaining output fields (here in the sample, number of delivered messages for message exchange status 'Done' and 'Error')
!petals-container-localtransporter-itf-dt-items.jpg|align=center!
# A new block *Custom data* appears, check *Use Per-Data Source Value* for all
!petals-container-localtransporter-itf-dt-custom.jpg|align=center!
# Create a new *Graph Template*:
!petals-container-localtransporter-itf-gt.jpg|align=center!
# Add a *Graph Item* to graphed one of the previous *data item* (here in the sample, the number of delivered messages for message exchange status 'Active')
!petals-container-localtransporter-itf-gt-active.jpg|align=center!
# Add a *Graph Item* to displayed the legend of your *data item* (here in the sample, the number of delivered messages for message exchange status 'Active')
!petals-container-localtransporter-itf-gt-active-legend.jpg|align=center!
# Add others *Graph Items* associated to the remaining *data items*
!petals-container-localtransporter-itf-gt-full.jpg|align=center!
# Go to the *data query* previoulsy created, and associate it to the just create graph template
!petals-container-localtransporter-itf-dq-associate.jpg|align=center!
# Associate outputs of the *data query* to the *graph items*
!petals-container-localtransporter-itf-dq-associate-output.jpg|align=center!
*Don't forget to check boxes.*
# Go to the *device* for which you want to create the graph, and *Create Graphs for this Host*
!petals-container-localtransporter-itf-ng.jpg|align=center!
Select yours endpoints for which you want create a graph

The new graph can be viewed after several minutes needed to collect data. By default a collect is done every 5 minutes. To view the graph, go into *Graph management"