Applying Transformations

1. Using Reduce Transformation

Property	Value	Description
type	`"reduce"`	Specifies the transformation type as reduce
includeTimeField	`false`	Whether to include the time field in the result
reducer	`"sum"`	The reduction function to apply (e.g., sum, avg, min, max)

Example: Applying Sum Reduction

{
  "type": "reduce",
  "includeTimeField": false,
  "reducer": "sum"
}

Note: The reduce transformation aggregates data points into a single value based on the specified reducer function.

2. Using Filter by Name

Property	Description	Values
type	Transformation type	`filterByName`
name	Name pattern to filter by	`string`
match	Matching strategy	`contains`, `equals`, `startsWith`, `endsWith`

Example: Filter metrics containing "cpu"

{
  "type": "filterByName",
  "name": "cpu",
  "match": "contains"
}

Note: The filterByName transformation is useful for reducing the number of time series displayed in a dashboard by filtering based on metric names.

3. Using Filter Data by Values

Option	Values	Description
type	`"filterByValue"`	Specifies the transformation type to filter data by value
settings	`{}`	Configuration options for the filter

Example: Filter Data by Value

{
  "type": "filterByValue",
  "settings": {
    "value": "error",
    "field": "level"
  }
}

Note: The field parameter specifies the field to filter on, and value is the value to match.

4. Using Merge Transformation

The Merge transformation in Grafana combines multiple data frames into a single data frame. It's particularly useful when you have multiple queries returning similar data and want to merge them into one view.

Option	Values	Description
mode	`merge`	Specifies the transformation type as merge
frameLimit	`100`	Maximum number of frames to merge (default: 100)

Example: Merging Query Results

{
  "refId": "A",
  "datasource": {
    "type": "prometheus",
    "uid": "prometheus-datasource-uid"
  },
  "expr": "up{job=\"api-server\"}"
}

Example: Merging with Different Time Series

{
  "refId": "B",
  "datasource": {
    "type": "prometheus",
    "uid": "prometheus-datasource-uid"
  },
  "expr": "up{job=\"auth-server\"}"
}

Note: The Merge transformation works best when queries return data with matching time series labels.

Warning: If the frames have different time ranges, the merged result may not align properly.

5. Using Join by Field

Option	Values	Description
joinByField	`string`	The field name to join on. Must exist in both data frames.
leftFrame	`string`	The identifier of the left data frame to join from.
rightFrame	`string`	The identifier of the right data frame to join with.
mode	`string`	Join mode: `inner`, `left`, or `full`.

Example: Joining Data Frames by Field

const leftFrame = {
  name: "cpu_usage",
  fields: [
    { name: "time", values: [1, 2, 3] },
    { name: "cpu", values: [75, 80, 85] }
  ]
};

const rightFrame = {
  name: "memory_usage",
  fields: [
    { name: "time", values: [1, 2, 3] },
    { name: "memory", values: [512, 520, 530] }
  ]
};

const joinTransformation = {
  type: "joinByField",
  joinByField: "time",
  leftFrame: "cpu_usage",
  rightFrame: "memory_usage",
  mode: "left"
};

// Resulting joined frame has fields: time, cpu, memory

Note: The joinByField transformation is useful for combining data from multiple queries based on a common field like timestamp or ID.

6. Using Add Field from Calculation

Property	Description	Example
fieldRef	Reference to the field to use in calculation	`fieldRef: 'value'`
mathExpression	Mathematical expression to calculate new field	`mathExpression: 'A * 2'`
newFieldKey	Name of the new field to be created	`newFieldKey: 'doubledValue'`

Example: Add Field from Calculation

{
  "type": "math",
  "id": "add-field-calculation",
  "options": {
    "fieldRef": "value",
    "mathExpression": "A * 2",
    "newFieldKey": "doubledValue"
  }
}

7. Using Organize Fields

Field	Description	Options	Notes
name	Name of the field to organize	`string`	Required field
label	Label to display for the field	`string`	Optional, defaults to name
index	Position in the field list	`number`	Optional, defaults to end of list
hide	Whether to hide the field from display	`true` or `false`	Optional, defaults to `false`
calcs	Calculations to perform on the field	`array`	Optional, can include `mean`, `sum`, etc.

Example: Organize Fields in a Panel

{
  "panels": [
    {
      "id": 1,
      "type": "graph",
      "fieldConfig": {
        "overrides": [
          {
            "matcher": {
              "id": "byName",
              "options": "value"
            },
            "properties": [
              {
                "id": "name",
                "value": "value"
              },
              {
                "id": "label",
                "value": "Value"
              },
              {
                "id": "index",
                "value": 0
              },
              {
                "id": "hide",
                "value": false
              }
            ]
          }
        ]
      }
    }
  ]
}

Note: The index property allows reordering fields in the panel display. Fields with lower index values appear first.

8. Using Convert Field Type

Field	Type	Description
field	`string`	Name of the field to convert
type	`string`	Target data type (e.g., `number`, `string`, `boolean`)
unit	`string`	Unit to convert to (e.g., `percent`, `bytes`)

Example: Convert field to number

{
  "type": "convertFieldType",
  "field": "cpu_usage",
  "type": "number"
}

Example: Convert field to string with unit

{
  "type": "convertFieldType",
  "field": "memory_bytes",
  "type": "string",
  "unit": "bytes"
}

9. Using Group by Transformation

Property	Description	Values	Notes
type	Transformation type identifier	`groupby`	Required field to specify transformation
settings	Configuration options for grouping	`{}`	Object containing groupBy and other options

Example: Grouping by a field in a time series

{
  "type": "groupby",
  "settings": {
    "groupBy": [
      {
        "type": "field",
        "value": "status"
      }
    ]
  }
}

Note: The groupby transformation is particularly useful for aggregating data in dashboards where you want to group metrics by a specific field like status, region, or user type.

10. Using Series to Rows

Option	Values	Description
type	`"seriesToRows"`	Transformation type to convert series to rows
id	`"seriesToRows"`	Unique identifier for the transformation

Example: Convert Series to Rows

{
  "type": "seriesToRows",
  "id": "seriesToRows"
}

Note: This transformation is useful for flattening time series data into a tabular format where each series becomes a row.

11. Using Rows to Fields

Option	Values	Description
type	`"rowsToFields"`	Transformation type to convert rows into fields
id	`"rowsToFields"`	Unique identifier for the transformation
name	`"Rows to Fields"`	Name displayed in the Grafana UI

Example: Transforming Rows to Fields

{
  "id": "rowsToFields",
  "name": "Rows to Fields",
  "type": "rowsToFields"
}

Note: This transformation is useful when your data source returns rows that need to be pivoted into fields for better visualization or analysis.

12. Chaining Multiple Transformations

Example: Chaining transformations in Grafana

{
  "targets": [
    {
      "refId": "A",
      "expr": "rate(http_requests_total[5m])"
    },
    {
      "refId": "B",
      "expr": "sum by (method) (A)"
    },
    {
      "refId": "C",
      "expr": "sort_desc(B)"
    }
  ]
}

Transformation	Description	Usage
Reduce	Aggregates data points using a specified function (e.g., sum, avg)	Use to consolidate time series data
Filter by Name	Filters time series based on name patterns or regex	Helps isolate specific metrics from a large set
Sort	Orders time series by value or name	Useful for consistent visualization order
Organize Fields	Reorders or removes fields from results	Improves data clarity and visualization layout

Note: Transformations are applied in the order they appear in the transformation list. Each step modifies the data before passing it to the next.