API Overuse Support for Fredhopper Customers

The following described explanation of how an unlikely request for API Overuse support query is processed on our end, and what you can do to prepare for your request.

It is not a common request, but sometimes we do get support tickets from our customers that have seen unexpected use of our APIs by their infrastructure and would like us to help them understand why.

We would like to remind you a couple of points to start with:

• Fredhopper front-end APIs are completely passive. As in, they cannot generate usage on behalf of your website. Your website infrastructure - more likely your back-end servers - will make a call to Fredhopper (this is called a “Query”) and, upon receiving a Query, Fredhopper will respond. With this interaction, your infrastructure is in complete control of when and how much it utilises our services.
• Fredhopper is integrated with a “server-to-server” method. This means when your website loads on a shopper’s browser, the browser is not making a direct call to Fredhopper. It instead communicates with your web server, and your server makes a call to Fredhopper. Thus, we do not have any direct visibility if a call was made for a real customer, or a bot crawling your website, or have any ability to make a decision to not make a call.
• It’s highly likely that your licence contract with us is based on the utilisation of our APIs, as in, how many API calls/Queries you are licensed to make within a certain duration. Please contact your Customer Service Representative for confirmation and details.

Preparation for making a Support request and what to expect

During your onboarding with Fredhopper, we have shared best-use practices for a front-end integration with our services with you. One of those is a strong recommendation for you to log every query your servers are making towards Fredhopper, which is a part of Traffic Monitoring Best Practices.

These logs are key for our investigation.

When you create a support ticket with our Customer Service regarding an investigation for unexpected API over-usage, or raise it with your Customer Success Manager, please provide the logs from your side for the timeframe that you want us to investigate. Please also add traffic figures your website has experienced during the same timeframe.

This will allow our teams to be able to cross-check the logs on our side to understand if there’s a discrepancy between what you have seen and logged on your end, and ours, so we can provide further guidance.

Only with these additional resources provided, would it be possible for us to spot any integration problems, or logging inaccuracies on either end.

Regarding bot traffic

As mentioned earlier, due to the “server-to-server” integration with Fredhopper, we treat and count all queries that are coming from your web servers the same as, due to our position and responsibility in the integration, we cannot know whether the query you are making us is for a bot or a valid customer.

However, we understand that this situation is not ideal. We also want our customers to benefit from every response we return per query and provide a better Product Discovery experience for their actual shoppers. To help, we have collated some recommendations on the Handling Automated Traffic page, and we encourage you to review it.

We hope the contents of the document provide guidance and help in identifying bot traffic before a query is made to Fredhopper and count against your allowance.

If you need any further support or have questions, please do not hesitate to reach out to your Customer Success Manager.