Skip to content

FieldConfig🔗

The data model used in Grafana, namely the data frame, is a columnar-oriented table structure that unifies both time series and table query results.

Each column within this structure is called a field. A field can represent a single time series or table column.

Field options allow you to change how the data is displayed in your visualizations.

Definition🔗

class FieldConfig:
    """
    The data model used in Grafana, namely the data frame, is a columnar-oriented table structure that unifies both time series and table query results.
    Each column within this structure is called a field. A field can represent a single time series or table column.
    Field options allow you to change how the data is displayed in your visualizations.
    """

    # The display value for this field.  This supports template variables blank is auto
    display_name: typing.Optional[str]
    # This can be used by data sources that return and explicit naming structure for values and labels
    # When this property is configured, this value is used rather than the default naming strategy.
    display_name_from_ds: typing.Optional[str]
    # Human readable field metadata
    description: typing.Optional[str]
    # An explicit path to the field in the datasource.  When the frame meta includes a path,
    # This will default to `${frame.meta.path}/${field.name}
    # 
    # When defined, this value can be used as an identifier within the datasource scope, and
    # may be used to update the results
    path: typing.Optional[str]
    # True if data source can write a value to the path. Auth/authz are supported separately
    writeable: typing.Optional[bool]
    # True if data source field supports ad-hoc filters
    filterable: typing.Optional[bool]
    # Unit a field should use. The unit you select is applied to all fields except time.
    # You can use the units ID availables in Grafana or a custom unit.
    # Available units in Grafana: https://github.com/grafana/grafana/blob/main/packages/grafana-data/src/valueFormats/categories.ts
    # As custom unit, you can use the following formats:
    # `suffix:<suffix>` for custom unit that should go after value.
    # `prefix:<prefix>` for custom unit that should go before value.
    # `time:<format>` For custom date time formats type for example `time:YYYY-MM-DD`.
    # `si:<base scale><unit characters>` for custom SI units. For example: `si: mF`. This one is a bit more advanced as you can specify both a unit and the source data scale. So if your source data is represented as milli (thousands of) something prefix the unit with that SI scale character.
    # `count:<unit>` for a custom count unit.
    # `currency:<unit>` for custom a currency unit.
    unit: typing.Optional[str]
    # Specify the number of decimals Grafana includes in the rendered value.
    # If you leave this field blank, Grafana automatically truncates the number of decimals based on the value.
    # For example 1.1234 will display as 1.12 and 100.456 will display as 100.
    # To display all decimals, set the unit to `String`.
    decimals: typing.Optional[float]
    # The minimum value used in percentage threshold calculations. Leave blank for auto calculation based on all series and fields.
    min_val: typing.Optional[float]
    # The maximum value used in percentage threshold calculations. Leave blank for auto calculation based on all series and fields.
    max_val: typing.Optional[float]
    # Convert input values into a display string
    mappings: typing.Optional[list[dashboard.ValueMapping]]
    # Map numeric values to states
    thresholds: typing.Optional[dashboard.ThresholdsConfig]
    # Panel color configuration
    color: typing.Optional[dashboard.FieldColor]
    # The behavior when clicking on a result
    links: typing.Optional[list[dashboard.DashboardLink]]
    # Alternative to empty string
    no_value: typing.Optional[str]
    # custom is specified by the FieldConfig field
    # in panel plugin schemas.
    custom: typing.Optional[object]

Methods🔗

to_json🔗

Converts this object into a representation that can easily be encoded to JSON.

def to_json() -> dict[str, object]

from_json🔗

Builds this object from a JSON-decoded dict.

@classmethod
def from_json(data: dict[str, typing.Any]) -> typing.Self