Hands-On Microservices with Kubernetes

Running distributed queries

We agree that each microservice should have its own data store. This means that the overall state of the system will be distributed across multiple data stores, accessible only from their own microservices. Most interesting queries will involve data available in multiple data stores. Each consumer could just access all these microservices and aggregate all the data to satisfy their query. However, that is sub-optimal for several reasons:

• Consumers are intimately aware of how data is managed by the system.
• Consumers need to get access to each and every service that stores data relevant to the query.
• Changing the architecture might require changes to a lot of consumers.

There are two common solutions to address ...