An external web UI application currently accepts occasional HTTP requests from client web browsers to change (insert, update, or delete) inventory pricing information in an inventory system's database. Each inventory pricing change must be transformed and then synchronized with multiple customer experience systems in near real-time (in under 10 seconds). New customer experience systems are expected to be added in the future.
The database is used heavily and limits the number of SELECT queries that can be made to the database to 10 requests per hour per user.
What is the most scalable, idiomatic (used for its intended purpose), decoupled. reusable, and maintainable integration mechanism available to synchronize each inventory pricing change with the various customer experience systems in near real-time?
- Write a Mule application with a Database On Table Row event source configured for the inventory pricing database, with the watermark attribute set to an appropriate database column
In the same now, use a Scatter-Gather to call each customer experience system's REST API with transformed inventory-pricing records - Add a trigger to the inventory-pricing database table so that for each change to the inventory pricing database, a stored procedure is called that makes a REST call to a Mule application
Write the Mule application to publish each Mule event as a message to an Anypoint MQ exchange Write other Mule applications to subscribe to the Anypoint MQ exchange, transform each received message, and then update the Mule application's corresponding customer experience system(s) - Replace the external web UI application with a Mule application to accept HTTP requests from client web browsers
In the same Mule application, use a Batch Job scope to test if the database request will succeed, aggregate pricing changes within a short time window, and then update both the inventory pricing database and each customer experience system using a Parallel For Each scope - Write a Mule application with a Database On Table Row event source configured for the inventory pricing database, with the ID attribute set to an appropriate database column
In the same flow, use a Batch Job scope to publish transformed Inventory-pricing records to an Anypoint MQ queue
Write other Mule applications to subscribe to the Anypoint MQ queue, transform each received message, and then update the Mule application's corresponding customer experience system(s)