minFraud Alerts

After initial scoring, we continue to monitor transactions with risk scores less than or equal to 10 for another 24 hours. If we receive new information related to these transactions that lead to a transaction having a re-calculated risk score greater than or equal to 75, we send out a minFraud Alert.

You can receive minFraud Alerts via email or webhook.

Receive minFraud Alerts via email

You can set an email where you would like to receive minFraud Alerts through the minFraud Alert configuration screen in your account portal (login required).

Receive minFraud Alerts via webhook

You can also receive minFraud Alerts by webhook, and integrate your own response to these alerts on your server. You can configure your minFraud Alert webhook settings through the minFraud Alert configuration screen in your account portal (login required).

Configuration

The alert webhook URL should be an HTTPS endpoint that accepts GET requests. The User-Agent of the minFraud Alert service is MaxMind MinFraud Alert Robot.

Sample minFraud Alert webhook request

1http://yourdomain/yoururl?i=24.24.24.24&maxmindID=1234ABCD&domain=sample.com&city=Anytown&region=CA&country=US&date=Jan.+1,+1970&txnID=foo123&reason=IP+address+has+been+marked+as+a+high-risk+IP&reason_code=HIGH_RISK_IP&minfraud_id=2afb0d26-e3b4-4624-8e66-fd10e64b95df&shop_id=shop321

Parameters

The following are the parameters that will be included in the query string of the request. Additional parameters may be added in the future.

Parameter	Description
`city`	The billing city included in the original minFraud request.
`country`	The billing country included in the original minFraud request.
`date`	The date of the original minFraud request, e.g., `Nov. 1, 2019`.
`domain`	The email domain included in the original minFraud request.
`i`	The IP address included in the original minFraud request.
`maxmindID`	The minFraud Legacy `maxmindID` of the original minFraud request.
`minfraud_id`	The minFraud ID of the original minFraud request.
`new_risk_score`	The updated risk score, calculated after your initial query with additional information.
`old_risk_score`	The risk score as originally calculated.
`postal`	The billing postal included in the original minFraud request.
`reason`	A human-readable description of why the minFraud alert was sent. See below for the current list of possible reasons.
`reason_code`	A fixed, machine-readable code for the reason why the minFraud alert was sent. See below for the current list of possible reasons.
`region`	The billing region included in the original minFraud request.
`shop_id`	The shop ID included in the original minFraud request. This will only be set if the original request included a shop ID.
`txnID`	The transaction ID included in the original minFraud request.
`updated_at`	The date and time at which the new risk score was calculated, in RFC3339 format, e.g., `2019-11-01T12:34:56Z`

Possible alert reasons

These are the possible reason_codes which might be returned with a minFraud Alert. These codes are subject to change.

`reason_code`	`reason`
`CARDER_EMAIL`	Email on order was flagged as high-risk email, as it was associated with another high-risk order
`HIGH_RISK_DEVICE`	Device ID has been marked as a high-risk device ID
`HIGH_RISK_IP`	IP address has been marked as a high-risk IP
`HOSTING_PROVIDER`	IP is from High Risk Hosting Provider
`MANUAL_REVIEW`	The transaction was flagged as risky after manual review
`POSTAL_VELOCITY`	IP address had high velocity of orders (e.g. different zipcodes on same IP address)

Securing your webhook

You can secure your webhook by verifying that alert requests come from MaxMind:

Check that requests come from one of the MaxMind IPs we list below.
Check that requests are signed using a secret you configure in your minFraud Alert configuration screen in your account portal. Learn more about setting up signed requests below.

Alert webhook request IPs

minFraud alert requests may come from one of the following IP addresses. This list is subject to change.

35.186.179.139

Signed requests

Alert requests can be signed by MaxMind. To enable signed requests, visit your minFraud Alert configuration screen in your account portal (login required) and set an Alert Webhook Secret.

Once you set a secret, each alert request will have an HTTP header with a signature of the request. You can calculate the signature of the request on your server and check it matches the signature in the header to verify the request is authentic.

For example, if the secret is supersekret-0123456789 and the alert request is:

1http://yourdomain/yoururl?i=24.24.24.24&maxmindID=1234ABCD&domain=sample.com&city=Anytown&region=CA&country=US&date=Jan.+1,+1970&txnID=foo123&reason=IP+address+has+been+marked+as+a+high-risk+IP&reason_code=HIGH_RISK_IP&minfraud_id=2afb0d26-e3b4-4624-8e66-fd10e64b95df&shop_id=shop321

Then the request will include this header:

X-MaxMind-Alert-HMAC-SHA256: c333974ffdddfa5d1ba866171847bacf1af06bbe6ff6b6ce73632116ed36f617

The signature is the hex encoded HMAC-SHA256 of the URL query parameters - everything after the ?. In this case, the text to be signed is:

1i=24.24.24.24&maxmindID=1234ABCD&domain=sample.com&city=Anytown&region=CA&country=US&date=Jan.+1,+1970&txnID=foo123&reason=IP+address+has+been+marked+as+a+high-risk+IP&reason_code=HIGH_RISK_IP&minfraud_id=2afb0d26-e3b4-4624-8e66-fd10e64b95df&shop_id=shop321

You can check the signaturing using something like the following code:

1require 'openssl'
2require 'openssl/hmac'
3require 'rack/utils'
4
5def verify_signature(secret, header, query)
6  expectedMAC = OpenSSL::HMAC.hexdigest('SHA256', secret, query)
7
8  raise 'invalid signature' if !Rack::Utils.secure_compare(header, expectedMAC)
9end

 1import (
 2	"crypto/hmac"
 3	"crypto/sha256"
 4	"encoding/hex"
 5)
 6
 7func verifySignature(secret, header, query string) (bool, error) {
 8  gotMAC, err := hex.DecodeString(header)
 9  if err != nil {
10    return false, err
11  }
12
13  mac := hmac.New(sha256.New, []byte(secret))
14  mac.Write([]byte(query))
15  expectedMAC := mac.Sum(nil)
16
17  return hmac.Equal(gotMAC, expectedMAC), nil
18}

To avoid timing attacks, please use a secure comparison function rather the equals operator of your language.