Request Validator

Looking for the plugin's configuration parameters? You can find them in the Request Validator configuration reference doc.

Validate requests before they reach their upstream service. This plugin supports validating the schema of the body and the parameters of the request using either Kong’s own schema validator (body only) or a JSON Schema Draft 4 compliant validator.

Content-Type validation

The request Content-Type header is validated against the plugin’s allowed_content_types setting. If the request Content-Type is not listed, the request will be rejected with an HTTP/400 error: {"message":"specified Content-Type is not allowed"}

The parameter is strictly validated, which means a request with a parameter (for example, application/json; charset=UTF-8) is NOT considered valid for one without the same parameter (for example, application/json). The type, subtype, parameter names, and the value of the charset parameter are not case sensitive based on the RFC explanation.

When setting this configuration, the Content-Type header only gets validated when body_schema is configured.

Body validation

Request body validation is only performed for requests with application/json Content-Type or when +json is suffixed in the request Content-Type’s subtype (for example, application/merge-patch+json).

For requests with any other allowed Content-Type, body validation is skipped. In that case, the request is proxied to the upstream without validating the body.

Examples

Overview

By applying the plugin to a service, all requests to that service will be validated before being proxied.

curl -i -X POST http://localhost:8001/services/{service}/plugins \
  --data "name=request-validator" \
  --data "config.version=kong" \
  --data 'config.body_schema=[{\"name\":{\"type\": \"string\", \"required\": true}}]'

plugins:
- name: request-validator
  service: {service}
  config:
    version: kong
    body_schema:
      name:
        type: string
        required: true

In this example, the request body data would have to be valid JSON and conform to the schema specified in body_schema - i.e., it would be required to contain a name field only, which needs to be a string.

In case the validation fails, a 400 Bad Request will be returned as the response.

Schema definition

For using the JSON Schema Draft 4-compliant validator, see the JSON Schema website for details on the format and examples. The rest of this paragraph will explain the Kong schema.

The config.body_schema field expects a JSON array with the definition of each field expected to be in the request body; for example:

Each field definition contains the following attributes:

Additionally, specific types may have their own required fields:

Map:

Example string-boolean map:

{
  "type": "map",
  "keys": {
    "type": "string"
  },
  "values": {
    "type": "boolean"
  }
}

Array:

Example integer array schema:

{
  "type": "array",
  "elements": {
    "type": "integer"
  }
}

Record:

Example record schema:

{
  "type": "record",
  "fields": [
    {
      "street": {
        "type": "string",
      }
    },
    {
      "zipcode": {
        "type": "string",
      }
    }
  ]
}

The type field assumes one the following values:

• string
• number
• integer
• boolean
• map
• array
• record

Each field specification may also contain validators, which perform specific validations:

Note: To learn more, see Lua patterns.

Semantic validation for the JSON Schema format attribute

This feature is only supported in JSON Schema Draft 4.

Structural validation alone may be insufficient to validate that an instance meets all the requirements of an application. The format keyword is defined to allow interoperable semantic validation for a fixed subset of values that are accurately described by authoritative resources, be they RFCs or other external specifications. The following attributes are available:

Note: The value of the format attribute must be a string.

Example date schema:

{
  "type": "string",
  "format": "date"
}

JSON Schema example

[
  {
    "name": {
      "type": "string",
      "required": true
    }
  },
  {
    "age": {
      "type": "integer",
      "required": true
    }
  },
  {
    "address": {
      "type": "record",
      "required": true,
      "fields": [
        {
          "street": {
            "type": "string",
            "required": true
          }
        },
        {
          "zipcode": {
            "type": "string",
            "required": true
          }
        }
      ]
    }
  },
  {
    "born": {
      "type": "string",
      "format": "date-time",
      "required": true
    }
  }
]

Such a schema would validate the following request body:

{
  "name": "Gruce The Great",
  "age": 4,
  "address": {
    "street": "251 Post St.",
    "zipcode": "94108"
  },
  "born": "2009-07-20T08:30:37.012Z"
}

Parameter schema definition

You can set up definitions for each parameter based on the OpenAPI Specification and the plugin will validate each parameter against it. For more information, see the OpenAPI specification or the OpenAPI examples.

Fixed fields

Examples

In this example, use the plugin to validate a request’s path parameter.

1. Add a service to Kong:

   curl -i -X POST http://localhost:8001/services \
     --data name=httpbin \
     --data url=https://httpbin.konghq.com
   

2. Add a route with named capture group:

   curl -i -X POST http://localhost:8001/services/httpbin/routes \
     --data paths="~/status/(?<status_code>\d%+)" \
     --data strip_path=false
   

3. Enable request-validator plugin to validate body and parameter:

    curl -i -X POST http://localhost:8001/services/httpbin/plugins \
      --header "Content-Type: application/json" \
      --data @parameter_schema.json
   

   Content of the file parameter_schema.json:

    {
      "name": "request-validator",
      "config": {
        "body_schema": "{\"properties\":{\"name\":{\"type\":\"string\"}},\"required\":[\"name\"]}",
        "version": "draft4",
        "parameter_schema": [
          {
            "name": "status_code",
            "in": "path",
            "required": true,
            "schema": "{\"type\": \"number\"}",
            "style": "simple",
            "explode": false
          }
        ]
      }
    }
   

4. In these step examples, validation ensures that status_code is a number and the body contains a parameter called name.

   For example, if you try to send a request to an existing route with the path /status/abc:

    curl -i -X POST \
    --url http://localhost:8000/status/abc \
    --header 'Content-Type: application/json' \
    --data '{ "name": "foo" }'
   

   The proxy request is blocked because the path has a non-numerical status code, which doesn’t conform to the schema:

    HTTP/1.1 400 Bad Request
    ...
   
    {"message":"request param doesn't conform to schema"}
   

   A proxy request with a numeric status code is allowed:

    curl -i -X POST \
    --url http://localhost:8000/status/123 \
    --header 'Content-Type: application/json' \
    --data '{ "name": "foo" }'
   

   Response:

    HTTP/1.1 200 OK
   

Further references

The Kong schema validation format is based on the plugin schemas. For more information, see the Kong plugin docs on storing custom entities.