Review

In the previous two sections, we've managed to create the core Event management system and integrated RabbitMQ messaging using Spring AMQP. For performing CRUD operations, we've utilized the Spring Data JPA project. In this section, we will integrate MongoDB and add error persistence.

Where am I?

Error Persistence

It might be odd why we need to record errors in the application. Isn't the default logging feature enough? That depends. But for the purpose of this tutorial, we are doing it to explore MongoDB integration. There are also advantages with this approach. We can easily query and analyze our application's errors and provide a statistical view.

It's interesting that logging is the first use case listed in the MongoDB site and statistical analysis as the last well-suited use case (see MongoDB Use Cases).

We'll divide this page into five sections:

C. Error Persistence
- C1. Domain
- C2. Service
- C3. Aspect
- C4. Controller
- C5. View

C1. Domain

The domain contains a simple ErrorLog class. At the class head we've added @Document, a Spring Data MongoDB annotation, for efficiency reasons, and @Id to mark the field id as the class's identifier. If you scrutinize the application's domain classes, notice the @Id annotation in the Event.java (earlier) is from the javax.persistence package; whereas the @Id in ErrorLog.java is from the org.springframework.data.annotation package.

ErrorLog.java

C2. Service

Normally before we start developing the service layer we have to create first the data access objects (DAOs) manually. It becomes tedious as you keep repeating the same implementation. Gladly, Spring is here to rescue us from this repetitive task with Spring Data project.

Spring Data's main purpose is to provide us ready-made data access objects for various database sources. This means all we need to do is write our service to contain the business logic. And since we're using a MongoDB, we will use the corresponding Spring Data project: Spring Data MongoDB. (To see a list of other Spring Data projects, visit http://www.springsource.org/spring-data).

IErrorLogRepository.java

That's it. We're done with the service. Wait, where's the service? We don't create one since we don't have custom business logic that needs to be wrapped in a service. All we're doing is retrieval of records.

C3. Aspect

Error persistence, at its core, is no different with logging. It's a crosscutting concern. And similar with how we implemented AMQP messaging, we will use an Aspect to add error persistence.

We have declared EventMongoAspect with two pointcut expressions: one for the service and another for the controller. The interceptService() method handles and persists service errors, while the interceptController() method handles controller errors. We don't persist controller errors; instead we provide a default value so that when returning JSP pages we're not propagating stack traces to the browser.

EventMongoAspect.java

C4. Controller

The error controller is a simple controller with the single purpose of serving the error page.

ErrorController.java

C5. View

Our view is a single JSP page containing a table and an AJAX request for retrieving records. The setup is similar with the event-page.jsp (see A4. View), except we don't have dialogs for adding, editing, or deleting records.

The getRecords() function is the AJAX request responsible for retrieval and filling-in of records to the table. Then it will convert the table to a DataTables. The addition of DataTables is mainly for aesthetic purposes (and some nifty features like sorting and searching).

error-page.jsp