How it works...

In this recipe, we implement a listener function that consumes upstream events and populates a materialized view that is used by a Backend For Frontend service. This function is a stream processor, such as the one we discussed in the Creating a stream processor recipe in Chapter 1, Getting Started with Cloud-Native. The function performs a filter for the desired events and then transforms the data in a map step to the desired materialized view. The materialized view is optimized to support the requirements of the query needed by the BFF. Only the minimum necessary data is stored, and the optimal database type is used.

In this recipe, the database type is a Cognito dataset. A Cognito dataset is a good choice for a materialized view when network availability is intermittent, and thus an offline-first approach is needed to synchronize data to a user's devices. The data must also be specific to a user so that it can be targeted to the user based on the user's identityId.  Due to the intermittent nature of connectivity, the asOf timestamp is included in the record so that the user can access the latency of the data.