Docs » Dashboards » Specifying Settings for All Charts in a Dashboard

Specifying Settings for All Charts in a Dashboard


Need some context? Creating, Sharing, and Protecting Dashboards

Want some guidelines? Better Dashboards


There are many situations where you will want to align all of the charts on a dashboard to the same filters, time range, or chart display resolution, such as:

  • Troubleshooting

    One of the first steps in root-cause analysis using time series data is slicing and dicing that data along a variety of dimensions, to help highlight patterns or trends. You can do this by using the appropriate dimension, property or tag to filter all of the data on a dashboard.

  • Efficiently creating similar dashboards

    You may want to replicate the monitoring setup you have in one environment into another. You can accomplish this by adding a dimension to the relevant metrics, then saving the dashboard.

  • Using the same dashboard to view metrics for an entire fleet as well as select subsets of it

    Some dashboards (for example, SignalFx’s out-of-the-box Amazon Web Services EC2 dashboard) show the total number of instances, instances by type, etc. for the entirety of the AWS account that has been integrated with SignalFx. You can use the same dashboard to see just the subset for a particular region or availability zone, for example, by applying the appropriate filter.

To apply settings to all charts on a dashboard, you specify overrides. An “Overrides” bar is shown at the top of every dashboard and chart. Any values specified here (whether in a chart or in a dashboard) override any individual values specified in every chart on the dashboard.

On all charts and dashboards, you can use the overrides bar to specify filters, time range and chart display resolution. The overrides bar might also contain one or more dashboard variables.

If you make changes in the overrides bar, you can use Save or Save as to save the dashboard with your changes. To revert to the previous settings, click Reset.

Filters

You can filter the charts in a dashboard based on a number of criteria. Filters can be dimensions, properties, or tags. Dimensions and properties are key-value pairs. This capability lets you create dashboards on demand that are populated with just the relevant subset of your data.

../_images/filter-db-01.png

When you place your cursor in the filter field, a dropdown list displays the available options. If you select a key (dimension or property), SignalFx adds a colon : to the key and shows you the list of available values to choose from for that key. You can also type sf_tags to find a list of tags relevant to the dashboard you are viewing or, if you know the tag you want to use, you can type it in directly.

SignalFx lets you specify a filter using a dimension for which you have not yet started sending data. Just type in the dimension and value you know you’ll want to filter by, and then press Enter. When data starts arriving for that dimension, the dashboard will be filtered as you specified.

In the illustration below, one filter has already been selected in the filter field.

../_images/filter-db-02.png

Note

As with all overrides, if you are viewing a chart in the Chart Builder when you specify a dashboard filter, it will be applied to all charts on the dashboard when you save and close the chart.

Any filters you apply here override any conflicting filters specified in plot lines.

Choosing what data to allow

If you hover over a filter setting, a caret appears at left to indicate that you have access to a dropdown menu. From here, you can specify how the filter should be applied, based on whether the property you specified applies to the incoming data.

../_images/filter-dropdown.png
  • Only allow data matching the filter condition (default)
    • Data that doesn’t match the condition will be excluded from the results.
    • Data that doesn’t contain the property will also be excluded from the results.
    • In this case, only data where the value of aws_region is us-east-1 will be included in the results.
  • Allow data matching the filter condition or missing <property>
    • If the incoming data contains the specified property, data that doesn’t match the condition will be excluded from the results.
    • If the incoming data doesn’t contain the specified property, the data will not be excluded from the results.
    • In this case, data where the value of aws_region is us-east-1, and also data that doesn’t contain the property aws_region, will be included in the results.

Filtering using multiple criteria (AND or OR)

You can specify multiple filters. If you select multiple tags, dimension keys or property keys, SignalFx will use the intersection of the selected filters, i.e. a Boolean AND.

You can also select multiple values for a single dimension or property key; in this case, SignalFx will use the union of the selected values (for the single key), i.e. a Boolean OR.

For example, the following two filters would be interpreted as:

(aws_region is us-east-1 OR aws_region is us-west-2) AND aws_instance_type is c3.2xlarge

../_images/multi-filter.png

Filtering by excluding specific criteria (creating a NOT filter)

You can also specify dimensions, properties or tags that you want to exclude from the results. To do so, type an exclamation point ! in the filter field, to represent a Boolean NOT. Once you’ve done this, the list of filtering options appears. By selecting an option from the list, you are saying that you want to exclude any items that match your selection.

Following up on the previous example, the following two filters would be interpreted as:

(aws_region is us-east-1 OR aws_region is us-west-2) AND aws_instance_type is NOT c3.2xlarge

../_images/multi-filter-not.png

Filtering from the data table

You can also add filters when you are viewing the data table for a chart. As you hover over dimensions in the data table, an Actions menu icon is displayed.

../_images/quick-filter-menu.png
  • To filter the view down so that only data with the specified value is shown, click the menu icon and select Filter. The corresponding filter is added in the overrides bar.

    ../_images/filter.png
  • To create a NOT filter so that data with the specified value is excluded from the view, click the menu icon and select Exclude. The exclamation point ! that represents a NOT filter is displayed in the filter token.

    ../_images/not-filter.png

Time range

Each chart within a dashboard will display on its x-axis the time range selected when creating the chart. (Single value charts and list charts show a timestamp rather than a range.) You can override the time range for all charts in a dashboard by specifying values in the Time Range selector, or by using your mouse to highlight a range in any of the charts.

As with all overrides, if you specify a time range when viewing the chart in the Chart Builder, it will be applied to all charts on the dashboard when you close the chart.

Any time range you specify here overrides any default time range you may have specified in the Chart Options tab.

Chart display resolution

When looking at a chart on a dashboard, every chart displays the time duration that each datapoint represents to the right of the chart title. You can use the chart resolution selector to increase or decrease the resolution. This feature is especially useful when different users are comparing two views of the same chart; users can use the resolution selector to ensure that they are viewing the same chart at the same resolution.

By default, charts are displayed in a dashboard in their default resolution. In the following illustration, all the charts are not displaying data at the same resolution.

../_images/chart-resolution-02.png

In the following illustration, the chart resolution setting for the dashboard was changed from Default to Very High. As you can see, all charts are now being displayed at 10s resolution.

../_images/chart-resolution-03.png

Notes

  • Setting a high display resolution doesn’t guarantee that all of the charts on the dashboard will have the same resolution. Some charts may already be at the finest possible resolution, in which case increasing the chart resolution setting for the dashboard will have no effect.
  • Single value and list charts always display at data resolution, and are unaffected by the chart resolution setting.
  • Chart resolution control only affects the visualization of charts, and does not affect detectors or alerts.
  • The chart resolution control is also available when viewing a chart in the Chart Builder. As with all overrides, if you specify a resolution when viewing the chart in the Chart Builder, it will be applied to all charts on the dashboard when you close the chart.

Max delay value

In addition to using the chart options tab to set a value for max delay for an individual chart, you can set a max delay value for a dashboard. As with all dashboard overrides, a dashboard max delay value applies to all the charts on the dashboard, overriding any individual chart’s max delay settings.

  1. From a dashboard’s Actions menu, select Dashboard Info.
  2. Click in the Max Delay field, then select an option from the dropdown or enter a value in seconds or minutes. You can enter a value up to 15 minutes, although values higher than 5 minutes are not recommended. Then click Save.

The selected setting will be displayed in the Overrides bar on the dashboard, and when someone opens a chart on the dashboard.

To remove the override:

  1. If a chart is open, close the chart to return to the dashboard.
  2. Click on the Max Delay value shown on the overrides bar to display the Dashboard Info tab.
  3. Click Reset to default, then click Save.

The max delay override option will no longer be displayed in the Overrides bar. Max delay values set for individual charts will be applied.

For more information on max delay, see Delayed datapoints.

Dashboard variables

Dashboard variables allow you to pin frequently used filters to the dashboard, removing the guesswork about which dimensions or properties to use. A dashboard can contain multiple dashboard variables. Filters applied through the use of dashboard variables are also retained by default when you drill down into individual charts, making troubleshooting workflows more efficient.

When you navigate among dashboards in a dashboard group, selections you make for dashboard variables in one dashboard are reflected in other dashboards which have the same dashboard variable. Thise feature helps you maintain context when moving from one dashboard to another in a dashboard group.

To add a variable to a dashboard, select Dashboard variables from the dashboard’s Actions menu.

Note

If you don’t see Dashboard variables, either don’t have write permissions for the dashboard you are viewing, or you are in a built-in dashboard.

You will see a modal window that allows you to create, edit or delete dashboard variables for the specific dashboard that you are viewing. If there are no dashboard variables currently in use, you will see that an Untitled variable has been added, with no values entered for any of its fields. The following sections describe how to use each of the fields.

../_images/dashboard-variables-01.png

Property

Select the dimension or property key that you want to use by clicking in the Property field and choosing from among the list of possible keys that appear in the dropdown menu.

Alias

Provide a label for the variable. The text you enter here will appear on the dashboard.

Default value and value required

Click in the Default value field to display a list of possible values in the dropdown menu. You can choose a dimension or property value to be used by default:

  • If you want to be able to use this dashboard without any filters (i.e. if the dashboard is useful without any filters applied), but want a specific filter to be applied to it when you arrive at the dashboard, add a value to the Default Value field and leave the box next to “Value for variable is required” unchecked.
  • If you want to be able to use this dashboard without any filters, and do not want any filter applied by default, leave the Default Value field empty and leave the box next to “Value for variable is required” unchecked.
  • If you want to always force a filter to be used with the dashboard, check the box by “Value for variable is required”. Specifying a Default Value is optional.

In this example, us-east-1 is the default value, and a value for this dashboard variable must always include be specified.

../_images/dashboard-variables-04.png

Description

If you add a description for the dashboard variable, it will appear in the mouseover text for the alias you’ve specified.

Suggestions and restricting suggestions

In some cases, you may want to ensure that certain property or dimension values are shown higher in the dropdown list on the dashboard. To select those values, click into the Suggestions field and select each of the values you want. Other values will still appear in the list, but below the ones you have specified. You can reorder the values by clicking and dragging the icon that appears on the left when you hover over a suggestion. In this example, us-east-1 and us-east-2 have been selected.

../_images/dashboard-variables-02.png

Instead of having other values appear below the ones you have suggested, you may want to have only your specified values appear in the dropdown list. To ensure this is the case, select and reorder the values you want in the Suggestions field as described above, and check “Restrict suggestions to predefined list”. The resulting list will include only your suggested values.

Apply to

You can choose whether the variable condition is applied to all plots on all charts on the dashboard.

  • All plots on all charts (default)

    The dashboard variable setting is applied as a filter to all plots on all charts, and behaves just like the Filter option for a dashboard. The dashboard will display data only for metrics that send in the specified property and meet the filter condition.

    Common use case for using this setting: You want to focus only on data that meets the variable condition, so you want to hide all other data.

    Possible side effect of using this setting: Because you are hiding data, your dashboard may have a number of blank charts.

  • Only plots with filters containing <property>

    The dashboard variable setting applies only to plots that have a filter for the property. Just like the Filter option, if a plot is filtered on a value for aws_availability_zone, a dashboard variable set to aws_availability_zone:US-east-1 will override that filter in the chart’s display.

    However, unlike the Filter option, the variable won’t apply to plots that don’t have a filter for the property; these will be displayed on the dashboard as if there were no filter applied.

    Common use case for using this setting: Suppose you have two charts, each with a plot that shows the mean of a metric that has the aws_availability_zone property. You want one chart to always show the mean aggregated across all zones, while in the other chart you want to be able to use a dashboard variable to drill down to a single zone. In the latter chart, you should add a filter for aws_availability_zone:* (or an OR filter that contains multiple zone values). When the dashboard variable is applied, the filter in that plot will be “replaced” by the specified availability zone, so the chart will display values only for that zone. However, because no filter has been applied to the plot in the first chart, the mean across all zones will still be displayed.

When you return to the dashboard and hover over the dashboard variable field, you will either see a tooltip telling you the variable applies to all charts, or you will see a different message along with some charts being highlighted on the dashboard. In the latter case, the dashboard variable is being applied to those charts because at least one plot in the chart has a filter matching the dashboard variable property.

../_images/dashboard-variables-05.png

Allow data

You can choose whether the variable condition displays data that is not sending the property specified in the filter. Assuming the dashboard variable on the dashboard is set to use the default value shown above (aws_region is us-east-1), this option is implemented as follows.

  • Only allow data matching the filter condition (default)
    • Data that doesn’t match the condition will be excluded from the results.
    • Data that doesn’t contain the property will also be excluded from the results.
    • In this case, only data where the value of aws_region is us-east-1 will be included in the results.
  • Allow data matching the filter condition or missing <property>
    • If the incoming data contains the specified property, data that doesn’t match the condition will be excluded from the results.
    • If the incoming data doesn’t contain the specified property, the data will not be excluded from the results.
    • In this case, data where the value of aws_region is us-east-1, and also data that doesn’t contain the property aws_region, will be included in the results.

When you return to the dashboard and hover over the dashboard variable field, you will see a tooltip telling you what data is being displayed, based on what data you have allowed.

Note that this setting applies to whatever data is displayed based on the option you specified for Apply to. In other words, Apply to specifies how broadly the filter should be applied to the charts, while Allow data specifies the set of data to which the filter will apply.

Saving and using dashboard variables

When you have made all your changes, click Save to return to the dashboard. In the example below, we’ve specified a default value for the variable, so it is applied to the dashboard immediately.

../_images/dashboard-variables-03.png

To change the value of the filter applied, click on the tokenized filter name to edit it, or on the x to delete it and select from the list of possible values. If the variable is required, clicking x won’t delete the filter; you will have to specify a value.