Skip to content

Latest commit

 

History

History
 
 

recaptcha-enterprise

reCAPTCHA Enterprise & Apigee X/hybrid

Google reCAPTCHA Enterprise can be leveraged by Apigee in order to protect backend systems from fraudulent API clients in the form of bots.

A reCAPTCHA token is silently and periodically retrieved by a client app and transmitted to an Apigee runtime when an API is invoked.

How it works?

The sharedflow configuration - that is proposed through this Apigee devrel reference - is able to get a reCAPTCHA token validation status and a risk score from the Google reCAPTCHA Enterprise assessment endpoint. The risk score is a decimal value between 0.0 and 1.0.

The score 1.0 indicates that the interaction poses low risk and is very likely legitimate, whereas 0.0 indicates that the interaction poses high risk and might be fraudulent. Between both extremes, the sharedflow's processing decides if an API invocation must be rejected or not. For the purpose of this reference we consider a minimum score of 0.6: this value is configurable and can be set to a higher or lower value depending on the risk profile of the client application.

Apigee runtime options

The reCAPTCHA enterprise reference can be deployed on both Apigee X and hybrid. This reference would also work with Apigee Edge if the Service Account token is obtained through the sharedflow, which invokes the Google reCAPTCHA enterprise endpoint (sf-recaptcha-enterprise-v1).

Dependencies

This reference leverages Google reCAPTCHA Enterprise. Therefore, it is important to note that:

  • The reCAPTCHA Enterprise API (recaptchaenterprise.googleapis.com) must be enabled in your GCP project.
  • A GCP service account is needed by the Apigee configuration to securely invoke the Google reCAPTCHA Enterprise assessment endpoint. This service account is created during the deployment process on the GCP project you are currently using: the pipeline.sh will attempt to create a service account only if it doesn't exist.

In case you want to create this service account manually - or with Terraform, please note that the role roles/recaptchaenterprise.agent must be granted to it.

Quick start

Apigee X / hybrid

export APIGEE_X_ORG=xxx
export APIGEE_X_ENV=xxx
export APIGEE_X_HOSTNAME=api.example.com

./pipeline.sh

Deployment options

There are 2 options to deploy the reCAPTCHA enterprise reference.

Option 1 is the default deployment option.

Option 1: reCAPTCHA Enterprise is mocked

export IS_RECAPTCHA_MOCK_ENABLED=true

With this option (default) the reCAPTCHA enterprise assessment is not invoked but a mock response similar to a real one is delivered. In this case, the risk score is based on the reCAPTCHA token that is transmitted through the X-Recaptcha-Token HTTP header.

Here are the possible values for the mock reCAPTCHA token, the related risk score and token validity:

Recaptcha-Token Risk Score is Token Valid?
X-RECAPTCHA-TOKEN-0 0 yes
X-RECAPTCHA-TOKEN-0.1 0.1 yes
X-RECAPTCHA-TOKEN-0.2 0.2 yes
X-RECAPTCHA-TOKEN-0.3 0.3 yes
X-RECAPTCHA-TOKEN-0.4 0.4 yes
X-RECAPTCHA-TOKEN-0.5 0.5 yes
X-RECAPTCHA-TOKEN-0.6 0.6 yes
X-RECAPTCHA-TOKEN-0.7 0.7 yes
X-RECAPTCHA-TOKEN-0.8 0.8 yes
X-RECAPTCHA-TOKEN-0.9 0.9 yes
X-RECAPTCHA-TOKEN-1 1 yes
Any other value NA no

Functional tests are executed only when this deployment option is selected.

Option 2: reCAPTCHA Enterprise is used

export IS_RECAPTCHA_MOCK_ENABLED=false

With this option, it is not possible to execute functional tests. Nevertheless, 2 client applications are created on Apigee.

A reCAPTCHA sitekey is created for each application:

  • Always 1 is a sitekey that will always deliver a reCAPTCHA token associated to a risk score of 1 (human)
  • Always 0 is a sitekey that will always deliver a reCAPTCHA token associated to a risk score of 0 (bot)

In order to deliver a valid reCAPTCHA token for test, an API proxy (recaptcha-deliver-token-v1) is deployed when this option is selected. The intent of this proxy is to deliver a simple HTML page, which executes a Javascript client code that can retrieve a valid reCAPTCHA token, based on one of the 2 sitekeys created on your GCP platform during the deployment process.

You can invoke the recaptcha-deliver-token-v1 proxy using your favorite web browser using an URL of this form:

https://${APIGEE_X_HOSTNAME}/recaptcha/v1/token?sitekey=<your-sitekey-here>

As a result on the Web page, you get a valid reCAPTCHA Enterprise token that you can copy/paste for testing purposes:

export RECAPTCHA_TOKEN='<paste-the-token-here>'

Script outputs

The pipeline script deploys on Apigee X or hybrid a sharedflow (sf-recaptcha-enterprise-v1) containing the full configuration of the reCAPTCHA enterprise reference as well as the following artifacts:

  • recaptcha-data-proxy-v1: a data proxy, which calls the reCAPTCHA enterprise sharedflow. The target endpoint of this proxy is httpbin.org

  • recaptcha-deliver-token-v1: an API proxy used to deliver an HTML page that includes a valid reCAPTCHA token (cf. Option 2 above). This proxy is not intended to be used in production but only during test phases.

  • The RecaptchaEnterprise API product

  • A developer (Jane Doe)

  • app-recaptcha-enterprise: a single developer app when Option 1 has been selected

  • 2 developer apps with real app credentials and reCAPTCHA enterprise sitekeys when Option 2 has been selected:

    • app-recaptcha-enterprise-always0
    • app-recaptcha-enterprise-always1

Creating Custom Reports for Token Validity and Risk Score

The sharedflow sf-recaptcha-enterprise-v1 contains 2 DataCapture policies, which are used to retrieve the token validity (true or false) and risk score.

These values are extracted from the Google reCAPTCHA enterprise assessment response and store into 2 data collectors:

  • dc_riskScore
  • dc_TokenValidity

From there, it is possible to create custom reports. Here is an example for each data collector:

Custom Report for Risk Score

Custom Report Risk Score

The cURL command to create this report is the following:

$ export APIGEE_TOKEN=$(gcloud auth print-access-token)

$ curl -H "Content-Type: application/json" \
-H "Authorization: Bearer $APIGEE_TOKEN" \
--data '{"name":"my-riskscore-report",
"displayName":"my-riskscore-report",
"metrics":[{"name":"dc_riskScore",
"function":"avg"},
{"name":"dc_riskScore","function":"min"},
{"name":"dc_riskScore","function":"max"},
{"name":"message_count","function":"sum"}],
"dimensions":["apiproxy","proxy_basepath","proxy_pathsuffix"],
"chartType":"column"}' \
https://apigee.googleapis.com/v1/organizations/"$APIGEE_X_ORG"/reports

Custom Report for Token Validity

Custom Report Token Validity

The Client IP address (X-Forwarded-For HTTP header) has been intentionally hidden on this report.

The cURL command to create this report is the following:

$ export APIGEE_TOKEN=$(gcloud auth print-access-token)

$ curl -H "Content-Type: application/json" \
-H "Authorization: Bearer $APIGEE_TOKEN" \
--data '{"name":"my-token-validity-report",
"displayName":"my-token-validity-report",
"metrics":[{"name":"message_count",
"function":"sum"}],
"dimensions":[
"dc_tokenValidity",
"apiproxy",
"proxy_basepath",
"proxy_pathsuffix",
"client_id",
"x_forwarded_for_ip"],
"chartType":"column"}' \
https://apigee.googleapis.com/v1/organizations/"$APIGEE_X_ORG"/reports

reCAPTCHA Enterprise & Apigee Sequence Diagram

The following sequence diagram provides all the interactions between:

  • End-user
  • Client app
  • Apigee: sharedflow and data proxy
  • Google reCAPTCHA enterprise endpoint
  • Backend

This sequence diagram is available as a text file.

If needed, you can modify this file and re-generate the related picture (png) using the following command:

./generate_docs.sh

Here is the original sequence diagram:

reCAPTCHA Enterprise

Testing the reCAPTCHA Enterprise reference

In order to test this reference, you need an HTTP client.

cURL command

Using cURL, the request is the following:

curl -H "X-Apikey: <consumer-key>" \
     -H "X-Recaptcha-Token: <recaptcha-token>" \
     https://${APIGEE_X_HOSTNAME}/recaptcha/v1/data/headers

Token validity

A reCAPTCHA enterprise token has a validity period of 2 minutes and can only be used once to avoid replay attacks. If a token is reused, it will be considered as invalid with no score associated to it.