JSON Parser Version 2 Plugin

This parser takes valid JSON input and turns it into line protocol. The query syntax supported is GJSON Path Syntax, you can go to the GJSON playground to test out your GJSON path expressions. You can find multiple examples under the testdata folder.

[!WARNING] In the current state of the implementation, the json_v2 parser should be avoided in favor of the XPath parser, especially when working with arrays.

Configuration

 [[inputs.file]]
    urls = []
    data_format = "json_v2"
    [[inputs.file.json_v2]]
        measurement_name = "" # A string that will become the new measurement name
        measurement_name_path = "" # A string with valid GJSON path syntax, will override measurement_name
        timestamp_path = "" # A string with valid GJSON path syntax to a valid timestamp (single value)
        timestamp_format = "" # A string with a valid timestamp format (see below for possible values)
        timestamp_timezone = "" # A string with a valid timezone (see below for possible values)
        [[inputs.file.json_v2.tag]]
            path = "" # A string with valid GJSON path syntax to a non-array/non-object value
            rename = "new name" # A string with a new name for the tag key
            ## Setting optional to true will suppress errors if the configured Path doesn't match the JSON
            optional = false
        [[inputs.file.json_v2.field]]
            path = "" # A string with valid GJSON path syntax to a non-array/non-object value
            rename = "new name" # A string with a new name for the tag key
            type = "int" # A string specifying the type (int,uint,float,string,bool)
            ## Setting optional to true will suppress errors if the configured Path doesn't match the JSON
            optional = false
        [[inputs.file.json_v2.object]]
            path = "" # A string with valid GJSON path syntax, can include array's and object's

            ## Setting optional to true will suppress errors if the configured Path doesn't match the JSON
            optional = false

            ## Configuration to define what JSON keys should be used as timestamps ##
            timestamp_key = "" # A JSON key (for a nested key, prepend the parent keys with underscores) to a valid timestamp
            timestamp_format = "" # A string with a valid timestamp format (see below for possible values)
            timestamp_timezone = "" # A string with a valid timezone (see below for possible values)

            ### Configuration to define what JSON keys should be included and how (field/tag) ###
            tags = [] # List of JSON keys (for a nested key, prepend the parent keys with underscores) to be a tag instead of a field, when adding a JSON key in this list you don't have to define it in the included_keys list
            included_keys = [] # List of JSON keys (for a nested key, prepend the parent keys with underscores) that should be only included in result
            excluded_keys = [] # List of JSON keys (for a nested key, prepend the parent keys with underscores) that shouldn't be included in result
            # When a tag/field sub-table is defined, they will be the only field/tag's along with any keys defined in the included_keys list.
            # If the resulting values aren't included in the object/array returned by the root object path, it won't be included.
            # You can define as many tag/field sub-tables as you want.
            [[inputs.file.json_v2.object.tag]]
                path = "" # # A string with valid GJSON path syntax, can include array's and object's
                rename = "new name" # A string with a new name for the tag key
            [[inputs.file.json_v2.object.field]]
                path = "" # # A string with valid GJSON path syntax, can include array's and object's
                rename = "new name" # A string with a new name for the tag key
                type = "int" # A string specifying the type (int,uint,float,string,bool)

            ### Configuration to modify the resulting line protocol ###
            disable_prepend_keys = false (or true, just not both)
            [inputs.file.json_v2.object.renames] # A map of JSON keys (for a nested key, prepend the parent keys with underscores) with a new name for the tag key
                key = "new name"
            [inputs.file.json_v2.object.fields] # A map of JSON keys (for a nested key, prepend the parent keys with underscores) with a type (int,uint,float,string,bool)
                key = "int"

You configure this parser by describing the line protocol you want by defining the fields and tags from the input. The configuration is divided into config sub-tables called field, tag, and object. In the example below you can see all the possible configuration keys you can define for each config table. In the sections that follow these configuration keys are defined in more detail.

General options

The optional measurement_name sets the name of emitted metrics to the provided string.

The optional measurement_name_path defines a GJSON query to set a metric name from the JSON input. The query must return a single data value otherwise the parser will use the default measurement name. This option takes precedence over measurement_name.

The optional timestamp_path defines a GJSON query to set the metric timestamp from the JSON input. The query must return a single data value otherwise the parser will default to the current time.

The optional timestamp_format is required in case timestamp_path is defined. The option can take the values unix, unix_ms, unix_us, unix_ns, or a Go "reference time" which is defined to be the specific time Mon Jan 2 15:04:05 MST 2006

The optional timestamp_timezone defaults to UTC but is required in case timestamp_path is defined. This option should be set to a Unix TZ value, such as America/New_York, to Local to utilize the system timezone or to UTC.

`field` and `tag` options

field and tag represent the elements of line protocol. You can use the field and tag config tables to gather a single value or an array of values that all share the same type and name. With this you can add a field or tag to a line protocol from data stored anywhere in your JSON. If you define the GJSON path to return a single value then you will get a single resulting line protocol that contains the field/tag. If you define the GJSON path to return an array of values, then each field/tag will be put into a separate line protocol (you use the # character to retrieve JSON arrays, see examples).

Note that objects are handled separately, therefore if you provide a path that returns a object it will be ignored. You will need use the object config table to parse objects, because field and tag doesn't handle relationships between data. Each field and tag you define is handled as a separate data point.

The notable difference between field and tag, is that tag values will always be type string while field can be multiple types. You can define the type of field to be any type that line protocol supports, which are:

float
int
uint
string
bool

field

Using this field configuration you can gather a non-array/non-object values. Note this acts as a global field when used with the object configuration, if you gather an array of values using object then the field gathered will be added to each resulting line protocol without acknowledging its location in the original JSON. This is defined in TOML as an array table using double brackets.

The required path setting contains a string with valid GJSON path syntax to a non-array/non-object value.

The optional name defines a string value to set the field name. If not defined it will use the trailing word from the provided query.

The optional type defines a string value to set the desired type (float, int, uint, string, bool). If not defined it won't enforce a type and default to using the original type defined in the JSON (bool, float, or string).

The optional optional setting can suppress errors if the configured path doesn't match the JSON data. This should be used with caution because it removes the safety net of verifying the provided path. An example case to use this is with the inputs.mqtt_consumer plugin when you are expecting multiple JSON files.

tag

Using this tag configuration you can gather a non-array/non-object values. Note this acts as a global tag when used with the object configuration, if you gather an array of values using object then the tag gathered will be added to each resulting line protocol without acknowledging its location in the original JSON. This is defined in TOML as an array table using double brackets.

The required path setting contains a string with valid GJSON path syntax to a non-array/non-object value.

The optional name defines a string value to set the field name. If not defined it will use the trailing word from the provided query.

For good examples in using field and tag you can reference the following example configs:

object

With the configuration section object, you can gather values from JSON objects. This is defined in TOML as an array table using double brackets.

The following keys can be set for `object`

The required path must define the GJSON path query that gathers the object.

Keys to define what JSON keys should be used as timestamps

The optional timestamp_key defines a JSON key (for a nested key, prepend the parent keys with underscores) for the value to be set as the timestamp from the JSON input.

The optional timestamp_format is required if timestamp_key is defined and must be set to unix, unix_ms, unix_us, unix_ns, or the Go "reference time" which is defined to be the specific time Mon Jan 2 15:04:05 MST 2006

The optional timestamp_timezone defaults to UTC but is required if timestamp_key is set. This option should be set to a Unix TZ value, such as America/New_York, to Local to utilize the system timezone or to UTC.

Configuration to define what JSON keys should be included and how (field/tag)

The optional included_keys define a list of key's that should be the only data included in the line protocol, by default it will include everything.

The optional excluded_keys define JSON keys to be excluded in the line protocol, for a nested key, prepend the parent keys with underscores

The optional tags define JSON keys to be set as tags instead of fields, if you define a key that is an array or object then all nested values will become a tag

The optional field defines an array table identical to the field table with two key differences. The path supports arrays and objects and is defined under the object table and therefore will adhere to how the JSON is structured. You want to use this if you want the field/tag to be added as it would if it were in the included_key list, but then use the GJSON path syntax.

The optional tag defines an array table identical to the tag table but with two key differences. The path supports arrays and objects and is defined under the object table and therefore will adhere to how the JSON is structured. You want to use this if you want the field/tag to be added as it would if it were in the included_key list, but then use the GJSON path syntax.

Configuration to modify the resulting line protocol

The optional disable_prepend_keys prevents resulting nested data to contain the parent key prepended to its key.

[!NOTE] Duplicate names can overwrite each other when this is enabled.

The optional renames table matches the JSON key with the desired name (opposed to defaulting to using the key), use names that include the prepended keys of its parent keys for nested results

The optional fields table matches the JSON key with the desired type (int, string, bool, float), if you define a key that is an array or object then all nested values will become that type

Arrays and Objects

The following describes the high-level approach when parsing arrays and objects:

Array: Every element in an array is treated as a separate line protocol

Object: Every key/value in a object is treated as a single line protocol

When handling nested arrays and objects, these above rules continue to apply as the parser creates line protocol. When an object has multiple array's as values, the array's will become separate line protocol containing only non-array values from the object. Below you can see an example of this behavior, with an input json containing an array of book objects that has a nested array of characters.

Example JSON:

{
    "book": {
        "title": "The Lord Of The Rings",
        "chapters": [
            "A Long-expected Party",
            "The Shadow of the Past"
        ],
        "author": "Tolkien",
        "characters": [
            {
                "name": "Bilbo",
                "species": "hobbit"
            },
            {
                "name": "Frodo",
                "species": "hobbit"
            }
        ],
        "random": [
            1,
            2
        ]
    }
}

Example configuration:

[[inputs.file]]
    files = ["./testdata/multiple_arrays_in_object/input.json"]
    data_format = "json_v2"
    [[inputs.file.json_v2]]
        [[inputs.file.json_v2.object]]
            path = "book"
            tags = ["title"]
            disable_prepend_keys = true

Expected line protocol:

file,title=The\ Lord\ Of\ The\ Rings author="Tolkien",chapters="A Long-expected Party"
file,title=The\ Lord\ Of\ The\ Rings author="Tolkien",chapters="The Shadow of the Past"
file,title=The\ Lord\ Of\ The\ Rings author="Tolkien",name="Bilbo",species="hobbit"
file,title=The\ Lord\ Of\ The\ Rings author="Tolkien",name="Frodo",species="hobbit"
file,title=The\ Lord\ Of\ The\ Rings author="Tolkien",random=1
file,title=The\ Lord\ Of\ The\ Rings author="Tolkien",random=2

You can find more complicated examples under the testdata folder.

Types

For each field you have the option to define the types. The following rules are in place for this configuration:

If a type is explicitly defined, the parser will enforce this type and convert the data to the defined type if possible. If the type can't be converted then the parser will fail.
If a type isn't defined, the parser will use the default type defined in the JSON (int, float, string)

The type values you can set:

int, bool, floats or strings (with valid numbers) can be converted to int
uint, bool, floats or strings (with valid numbers) can be converted to uint
string, any data can be formatted as a string
float, string values (with valid numbers) or integers can be converted to float
bool, the string values "true" or "false" (regardless of capitalization) or the integer values 0 or 1 can be turned to a bool

# README