Create search-based visualizations with ds.search

The data source type ds.search is used for ad hoc searches that you create in the visual editor. If you want to use a saved search, see Use reports and saved searches with ds.savedSearch.

You can specify the time range for your data sources using either the Dashboard Studio UI or by directly updating the source code.

When editing a data source with the UI, you can choose to use the following:

  • Any time range input on the page.
  • A static time range. For example, the last seven days, regardless of other input selections.
  • The dashboard default.

When editing the data source in the source code, the ds.search data source is assigned a unique ID when you create a search string in the Configuration panel. You can find the associated source code stanza by locating the unique ID. The stanza contains the data source type and any options you have made available.

The only option that must return events is the query option, which is an SPL search.

For example, the following data source uses the search 

index=_internal | head 500

to return results that can be displayed by adding the ID of the data source, "ds_BCsearch", to a visualization module as its primary data source. 

"dataSources": {
        "BCsearch": {
            "type": "ds.search",
            "options": {
                "query": "index=_internal | head 500",
                "queryParameters": {
                    "earliest": "0",
                    "latest": ""
                }
            }
        }
    },

The following is an example of a visualization using ds_BCsearch as the primary dataSource . 

"visualizations": {
        "viz_Example": {
            "type": "splunk.table",
            "options": {},
            "dataSources": {
                "primary": "BCsearch"
            }
        }
    },

The visualization must also appear in the layout. The following is a complete dashboard example using the two stanzas above. 

{
    "title": "Example",
    "description": "",
    "inputs": {},
    "defaults": {
        "visualizations": {
            "global": {
                "showLastUpdated": true
            }
        }
    },
    "visualizations": {
        "viz_Example": {
            "dataSources": {
                "primary": "BCsearch"
            },
            "options": {},
            "type": "splunk.table"
        }
    },
    "dataSources": {
        "BCsearch": {
            "options": {
                "query": "index=_internal | head 500",
                "queryParameters": {
                    "earliest": "0",
                    "latest": ""
                }
            },
            "type": "ds.search"
        }
    },
    "layout": {
        "globalInputs": [],
        "layoutDefinitions": {
            "layout_1": {
                "options": {
                    "width": 1801
                },
                "structure": [
                    {
                        "item": "viz_Example",
                        "position": {
                            "h": 600,
                            "w": 1270,
                            "x": 10,
                            "y": 50
                        },
                        "type": "block"
                    }
                ],
                "type": "absolute"
            }
        },
        "tabs": {
            "items": [
                {
                    "label": "New tab",
                    "layoutId": "layout_1"
                }
            ]
        }
    }
}

The resulting table visualization looks similar to this:

A table visualization using a search based data source.

Note: If you are using lookups in your search, make sure the lookup permissions are set to "global".

Modify the time range for a data source

By default all data sources that power visualizations on a dashboard are controlled by the default Global Time Range Picker. You can use the source editor to modify set time ranges for individual data source stanzas and override the Global Time Range Picker. Because it's an input, you can also delete or modify the default Global Time Range Picker in the source editor. To learn how to use data source options that affect the time range, for example, queryParameters, see Data source options. To read more about the time range picker, see Time range.

Set the event sampling ratio

By default, a Splunk search retrieves all events. However in some situations you might want to retrieve a sample set of events, instead of retrieving the entire event set, such as to determine the characteristics of a large data set without processing every event. See Event sampling.

The default sample ratio for a data source is 1, meaning that the search retrieves all events with no sampling applied. You can specify a sample ratio using the Event Sampling option. For example, if the sample ratio value is 100, each event has a 1 in 100 chance of being included in the result set.

You can also configure the event sample ratio in the source code using the sampleRatio option. The following shows the source code for a data source where the event sampling ratio is 1:100.
{
    "name": "Model usage by team",
    "options": {
        "query": "index=\"buttercup_enterprises_ai\" | stats count by team model_name | where count > 0  | table model_name team count | sort -value",
        "queryParameters": {
            "sampleRatio": "100"
        }
    },
    "type": "ds.search"
}

Limitations to search-based dashboards

Dashboards with many searches and exceeding 7,000 characters in the Uniform Resource Identifier (URI) do not consistently load and might remain in a "waiting on data" state. A large URI might exceed the limit of various server software used with the Splunk platform, such as Nginx.