Metric Tile

1. Select the Build my own tab. This enables you to create a custom tile.

2. Click the + QUERY and select the options from the following drop-down lists as you build the query:
   • Metrics - Select any metric that is being sent to OpsRamp or an OpsRamp-generated metric.
   • Filters - Select to filter the metrics based on the tags or labels that are present on those metrics.

Resource group

• Filter metrics based on a resource group.

  While building the query:

• Select the metric and click +Filters.

• Select groups from the list of tags.
  Note: You can also search using the search box.
• Select the operator and the value. The metric tile shows the metric details based on the resource group.
  
• You can also use a dashboard variable as a value.

Note: A resource can be a part of one group or multiple groups. These groups appear as a comma-separated list in the metric series.

Sites

• Filter metrics based on a Site.
  

  While building the query:
  

• Select the metric and click +Filters.

• Select sites from the list of tags. You can also search using the search box. Select the operator and the value. The metric tile shows the metric details based on the site.
  

• You can also use a dashboard variable as a value.

Service Map name

Filter a metric by the service map name.

• While building the query:
  • Select the metric and click +Filters.

• Select servicegroups from the list of tags. You can also search using the search box. Select the operator and the value. The metric tile shows the metric details based on the service map.
• If a resource is added in two service maps, then the service groups data should be seen in comma-separated values.
• You can also use a dashboard variable as a value.

Custom Attributes

Use custom attributes when querying metrics.

• Define custom attributes on the resources. Use these custom attributes to graph metric data.



Example: View the CPU, memory, disk metrics for a specific group of resources that are tagged with specific custom attributes.

• While building the query:
• Select the metric and click +Filters.



• Select the custom attribute, the operator and the value from the list of tags.

  Note: You can also search using the search box.

• The metric tile displays the metric series for all the resources that are tagged with the custom attributes for the specific metric.

Note: When creating a custom attribute value (in the Setup > Custom Attributes page), select the Metric Label checkbox. Only then the custom attribute will be considered as a label for the resource metrics.



• It takes 24 hours for the custom attribute labels to be displayed on the metric series.

• If you do not select the Metric Label checkbox while creating a custom attribute value, it will not be visible when you edit the custom attribute value.

• Metric label enabled custom attributes have limits. A maximum of 5 metric labels can be enabled per device.

• The metric label enabled custom attributes are displayed in the resource details page, in the Attributes section. This will allow you to decide which of the custom attributes you can consider as a label for the metrics.
  

Click here for more information on Custom Attributes.

• Operations - You can select a maximum of two operations ranging from aggregations such as average or summation to mathematical functions like logarithms or absolute values. You can use min, max, avg, topk for the series generated.
  The preview graph is displayed as you build the query.

• Advance Query - click </> to use the advanced query builder. Click </> again to close the advanced query field and switch to the basic query builder format.
  Note: If you make changes in the advanced query builder, you cannot switch back to the basic query builder.

  • Start typing a query and choose either a function or metric from the dropdown that appears, if you are using the advanced query option.
    See QUERYING PROMETHEUS to learn about PromQL queries.
  • Y-Axis Scaling: Y-axis dynamically adjusts when multiple queries are added, enabling effective comparison between queries with different ranges of values.
    The following is the Y-Axis Scaling visualization with the inputs:
    • Metric A = acs_rds_MySQL_ActiveSessions
    • Metric B = acs_rds_MySQL_ibufRequestW



• Forecast Graph: The forecast graph provides flexibility for you to specify a Prometheus query expression, enabling you to predict and visualize the forecasted value of time series data for the given metric.

  • You can also seamlessly navigate from any metric tile, with the specific time frame and relevant tags, to the log explorer simplifying the task of correlating metric data with log data, ensuring a streamlined and contextually relevant experience.

    1. Hover over the graph and click on any point, the VIEW LOGS is displayed.
    2. Click VIEW LOGS.
       Note: click the VIEW LOGS again to close the drilldown.

  

  Note: A maximum of 10 days can be predicted.

  Below is an example query using promql expression to display the forecast of a metric.

  Metric A: predict_linear(system_disk_utilization{host="device1"}[5d], 30*24*60*60)
  In the above example,

PROMQL Value	Description
predict_linear	promql expression to forecast the metric data.
system_disk_utilization	metricName for which the time series forecasting is to be visualized.
host	Device name for which the forecast is to be visualized
"5d"	Forecasting future window to predict and display the metric data for the next 5 days. Users can define based on the required prediction time period.
30*24*60*60	Learning time window, to look back 30 days of data to generate the forecasting analysis trend.

• Change Detection: The change detection provides flexibility in visualizing change detection patterns. You can now define the below expressions in your Prometheus query to generate and visualize the change detected occurrences.
• Upper limit.
• Actual Value.
• Lower Limit.
Below is an example of how these expressions are used to show change detection in the metric tile:

UpperLimit: avg_over_time(system_cpu_utilization{host=“QA-WIN12HYD-KL01”, instance=“CPU”}[4h])) + 3*(stddev_over_time(system_cpu_utilization{host=“QA-WIN12HYD-KL01”, instance=“CPU”}[4h]))


Actual Value: system_cpu_utilization{host=“QA-WIN12HYD-KL01”, instance=“CPU”}

For LowerLimit: avg_over_time(system_cpu_utilization{host=“QA-WIN12HYD-KL01”, instance=“CPU”}[4h])) - 3*(stddev_over_time(system_cpu_utilization{host=“QA-WIN12HYD-KL01”, instance=“CPU”}[4h]))


In the above example,

PROMQL Value	Description
"system_cpu_utilization"	metricName.
"4h"	The learning duration, which users can customize between 1h to 24h based on the requirements.
"3"	Sensitivity factor which is the multiple of standard deviations from the mean value, users can give any number but the recommended value would be below 10.

4. Click the optional Legend icon and enter {{ in the Query a Legend field to see a list of options.
   
5. Select the data type from the following options:

6. Click SAVE to view the results based on your selection. Hover over the results on the preview graph to view them individually.
7. Select the Value option to view the results as a single value.
   Note: You can add upto three queries, however, you will no longer be able to create additional ones.

8. You can sort the queries in ascending or descending order. This option is available for the List visualization.

9. Select the Instant toggle to view the recent value for a given metric.

   Note: The Instant option is available only for List, Gauge, and Honeycomb chart types.

10. Click the Last Hour to select the duration for which the data needs to be retrieved. The duration can be selected for:

    • Last Hour
    • Last 4 Hours
    • Last 8 Hours
    • Last 24 Hours
    • Last 7 Days
    • Last 15 Days
    • Last 30 Days
    • Month to Date

11. See Visualization to know on how to customize the Metric Tile.

12. Click Create.
    The Metric tile is created and displayed on the main dashboard page.

1. Select the Top Metric Utilization tab.

2. The default PROMQL query is already generated. Below is the PromQL:

topk(10,avg by (uuid,instance,name,__name__)({type="RESOURCE",performanceType="CPU"}))

Note: You also have the option to edit the PROMQL.

3. See Visualization to know on how to customize the Metric Tile.
4. Click Create.
   The Metric tile is created and displayed on the main dashboard page.

1. Select the Time Series tab.

2. The default PROMQL query is already generated. Below is the PromQL:

avg_over_time(system_cpu_utilization{type="RESOURCE"}[1h])

Note: You also have the option to edit the PROMQL.

3. See Visualization to know on how to customize the Metric Tile.
4. Click Create.
   The Metric tile is created and displayed on the main dashboard page.

1. Select the Metric Value tab.

2. The default PROMQL query is already generated. Below is the PromQL:

avg(last_over_time(system_cpu_utilization{type="RESOURCE"}[1h]))

Note: You also have the option to edit the PROMQL.

3. See Visualization to know on how to customize the Metric Tile.
4. Click Create.
   The Metric tile is created and displayed on the main dashboard page.

1. Select the Distribution by Metric tab.

2. The default PROMQL query is already generated. Below are the PromQL:

count((max({performanceType="CPU"}) by (uuid)) > 0 <= 40)

count((max({performanceType="CPU"}) by (uuid)) > 40 <= 80)

count((max({performanceType="CPU"}) by (uuid)) > 80 <= 100)

Note: You also have the option to edit the PROMQL.

3. See Visualization to know on how to customize the Metric Tile.
4. Click Create.
   The Metric tile is created and displayed on the main dashboard page.