In this post, I will show how we can use the Circuit Breaker pattern in a Spring Boot Application. When I say Circuit Breaker pattern, it is an architectural pattern. Netflix had published a library Hysterix for handling circuit breakers.  As part of this post, I will show how we can use a circuit breaker pattern using the resilence4j  library in a Spring Boot Application.

In other news, I recently released my book Simplifying Spring Security. If you are interested to learn about Spring Security, you can buy it here.

What is Circuit Breaker?

The concept of Circuit Breaker comes from Electrical Engineering. In most electricity networks, circuit breakers are switches that protect the network from damage caused by an overload of current or short circuits.

Similarly, in software, a circuit breaker stops the call to a remote service if we know the call to that remote service is either going to fail or time out. The advantage of this is to save resources and be proactive in our troubleshooting of the remote procedure calls.

The circuit breaker makes the decision of stopping the call based on the previous history of the calls. But there are alternative ways how it can handle the calls. Usually, it will keep track of previous calls. Suppose 4 out of 5 calls have failed or timed out, then the next call will fail. This helps to be more proactive in handling the errors with the calling service and the caller service can handle the response in a different way, allowing users to experience the application differently than an error page.

Another way a circuit breaker can act is if calls to remote service are failing in particular time duration.  A circuit breaker will open and will not allow the next call till remote service improves on error.

Resilience4J Library

We have our code which we call remote service. The circuit breaker module from resilience4j library will have a lambda expression for a call to remote service OR a supplier to retrieve values from the remote service call. I will show this as part of the example. The circuit breaker decorates this remote service call in such a way that it can keep track of responses and switch states.

Different configurations of Resilience4j Library

To understand the circuit breaker concept, we will look at different configurations this library offers.

slidingWindowType() – This configuration basically helps in making a decision on how the circuit breaker will operate. There are two types COUNT_BASED and TIME_BASED. COUNT_BASED circuit breaker sliding window will take into account the number of calls to remote service while TIME_BASED circuit breaker sliding window will take into account the calls to remote service in certain time duration.

failureRateThreshold() – This configures the failure rate threshold in percentage. If x percentage of calls are failing, then the circuit breaker will open.

slidingWindowSize() – This setting helps in deciding the number of calls to take into account when closing a circuit breaker.

slowCallRateThreshold() – This configures the slow call rate threshold in percentage. If x percentage of calls are slow, then the circuit breaker will open.

slowCallDurationThreshold – Time duration threshold about which calls are considered slow.

minimumNumberOfCalls() – A minimum number of calls required before which circuit breaker can calculate the error rate.

ignoreException() – This setting allows you to configure an exception that a circuit breaker can ignore and will not count towards the success or failure of a call of remote service.

waitDurationInOpenState() – Duration for which the circuit breaker should remain in the open state before transitioning into a half-open state.  The default value is 60 seconds.

Count-Based Circuit Breaker

While using resilience4j library, one can always use the default configurations that the circuit breaker offers. Default configurations are based on the COUNT-BASED sliding window type.

So how do we create a circuit breaker for the COUNT-BASED sliding window type?

      CircuitBreakerConfig circuitBreakerConfig = CircuitBreakerConfig.custom()
                .slidingWindowType(CircuitBreakerConfig.SlidingWindowType.COUNT_BASED)
                .slidingWindowSize(10)
                .slowCallRateThreshold(65.0f)
                .slowCallDurationThreshold(Duration.ofSeconds(3))
                .build();

        CircuitBreakerRegistry circuitBreakerRegistry =
                CircuitBreakerRegistry.of(circuitBreakerConfig);

        CircuitBreaker cb = circuitBreakerRegistry.circuitBreaker("BooksSearchServiceBasedOnCount");

In the above example, we are creating a circuit breaker configuration that includes a sliding window of type COUNT_BASED. This circuit breaker will record the outcome of 10 calls to switch the circuit-breaker to the closed state.  If 65 percent of calls are slow with slow being of a duration of more than 3 seconds, the circuit breaker will open.

CircuitBreakerRegistry is a factory to create a circuit breaker.

Time-Based Circuit Breaker

Now on Time-Based circuit breaker.

       CircuitBreakerConfig circuitBreakerConfig = CircuitBreakerConfig.custom()
                .slidingWindowType(CircuitBreakerConfig.SlidingWindowType.TIME_BASED)
                .minimumNumberOfCalls(3)
                .slidingWindowSize(10)
                .failureRateThreshold(70.0f)
                .build();

        CircuitBreakerRegistry circuitBreakerRegistry =
                CircuitBreakerRegistry.of(circuitBreakerConfig);

        CircuitBreaker cb = circuitBreakerRegistry.circuitBreaker("BookSearchServiceBasedOnTime");

In the above example, we are creating a circuit breaker configuration that includes a sliding window of type TIME_BASED. Circuit breaker will record the failure of calls after a minimum of 3 calls.  If 70 percent of calls fail, the circuit breaker will open.

Example of Circuit Breaker in Spring Boot Application

We have covered the required concepts about the circuit breaker. Now, I will show we can use a circuit breaker in a Spring Boot application.

On one side, we have a REST application BooksApplication that basically stores details of library books. On the other side, we have an application Circuitbreakerdemo that calls the REST application using RestTemplate. We will decorate our REST call through the circuit breaker.

BooksApplication stores information about books in a MySQL database table librarybooks. The REST Controller for this application has GET and POST methods.

package com.betterjavacode.books.controllers;

import com.betterjavacode.books.daos.BookDao;
import com.betterjavacode.books.models.Book;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.http.HttpStatus;
import org.springframework.http.ResponseEntity;
import org.springframework.web.bind.annotation.*;

import java.util.ArrayList;
import java.util.List;
import java.util.Optional;

@CrossOrigin("https://localhost:8443")
@RestController
@RequestMapping("/v1/library")
public class BookController
{
    @Autowired
    BookDao bookDao;

    @GetMapping("/books")
    public ResponseEntity<List> getAllBooks(@RequestParam(required = false) String bookTitle)
    {
        try
        {
            List listOfBooks = new ArrayList<>();
            if(bookTitle == null || bookTitle.isEmpty())
            {
                bookDao.findAll().forEach(listOfBooks::add);
            }
            else
            {
                bookDao.findByTitleContaining(bookTitle).forEach(listOfBooks::add);
            }

            if(listOfBooks.isEmpty())
            {
                return new ResponseEntity<>(HttpStatus.NO_CONTENT);
            }

            return new ResponseEntity<>(listOfBooks, HttpStatus.OK);
        }
        catch (Exception e)
        {
            return new ResponseEntity<>(null, HttpStatus.INTERNAL_SERVER_ERROR);
        }
    }

    @GetMapping("/books/{id}")
    public ResponseEntity getBookById(@PathVariable("id") long id)
    {
        try
        {
            Optional bookOptional = bookDao.findById(id);

            return new ResponseEntity<>(bookOptional.get(), HttpStatus.OK);
        }
        catch (Exception e)
        {
            return new ResponseEntity<>(null, HttpStatus.INTERNAL_SERVER_ERROR);
        }
    }

    @PostMapping("/books")
    public ResponseEntity addABookToLibrary(@RequestBody Book book)
    {
        try
        {
            Book createdBook = bookDao.save(new Book(book.getTitle(), book.getAuthor(),
                    book.getIsbn()));
            return new ResponseEntity<>(createdBook, HttpStatus.CREATED);
        }
        catch (Exception e)
        {
            return new ResponseEntity<>(null, HttpStatus.INTERNAL_SERVER_ERROR);
        }
    }

    @PutMapping("/books/{id}")
    public ResponseEntity updateABook(@PathVariable("id") long id, @RequestBody Book book)
    {
        Optional bookOptional = bookDao.findById(id);

        if(bookOptional.isPresent())
        {
            Book updatedBook = bookOptional.get();
            updatedBook.setTitle(book.getTitle());
            updatedBook.setAuthor(book.getAuthor());
            updatedBook.setIsbn(book.getIsbn());
            return new ResponseEntity<>(bookDao.save(updatedBook), HttpStatus.OK);
        }
        else
        {
            return new ResponseEntity<>(HttpStatus.NOT_FOUND);
        }
    }

    @DeleteMapping("/books/{id}")
    public ResponseEntity deleteABook(@PathVariable("id") long id)
    {
        try
        {
            bookDao.deleteById(id);
            return new ResponseEntity<>(HttpStatus.NO_CONTENT);
        }
        catch (Exception e)
        {
            return new ResponseEntity<>(HttpStatus.INTERNAL_SERVER_ERROR);
        }
    }
}

On the other side, our application Circuitbreakerdemo has a controller with thymeleaf template so a user can access the application in a browser.

For the demo purpose, I have defined CircuitBreaker in a separate bean that I will use in my service class.

    @Bean
    public CircuitBreaker countCircuitBreaker()
    {
        CircuitBreakerConfig circuitBreakerConfig = CircuitBreakerConfig.custom()
                .slidingWindowType(CircuitBreakerConfig.SlidingWindowType.COUNT_BASED)
                .slidingWindowSize(10)
                .slowCallRateThreshold(65.0f)
                .slowCallDurationThreshold(Duration.ofSeconds(3))
                .build();

        CircuitBreakerRegistry circuitBreakerRegistry =
                CircuitBreakerRegistry.of(circuitBreakerConfig);

        CircuitBreaker cb = circuitBreakerRegistry.circuitBreaker("BooksSearchServiceBasedOnCount");

        return cb;
    }

    @Bean
    public CircuitBreaker timeCircuitBreaker()
    {
        CircuitBreakerConfig circuitBreakerConfig = CircuitBreakerConfig.custom()
                .slidingWindowType(CircuitBreakerConfig.SlidingWindowType.TIME_BASED)
                .minimumNumberOfCalls(3)
                .slidingWindowSize(10)
                .failureRateThreshold(70.0f)
                .build();

        CircuitBreakerRegistry circuitBreakerRegistry =
                CircuitBreakerRegistry.of(circuitBreakerConfig);

        CircuitBreaker cb = circuitBreakerRegistry.circuitBreaker("BookSearchServiceBasedOnTime");
        return cb;
    }

I have defined two beans one for the count-based circuit breaker and another one for time-based.

The BookStoreService will contain a calling BooksApplication and show books that are available. This service will look like below:

@Controller
public class BookStoreService
{

    private static final Logger LOGGER = LoggerFactory.getLogger(BookStoreService.class);

    @Autowired
    public BookManager bookManager;

    @Autowired
    private CircuitBreaker countCircuitBreaker;

    @RequestMapping(value = "/home", method= RequestMethod.GET)
    public String home(HttpServletRequest request, Model model)
    {
        return "home";
    }

    @RequestMapping(value = "/books", method=RequestMethod.GET)
    public String books(HttpServletRequest request, Model model)
    {
        Supplier<List> booksSupplier =
                countCircuitBreaker.decorateSupplier(() -> bookManager.getAllBooksFromLibrary());

        LOGGER.info("Going to start calling the REST service with Circuit Breaker");
        List books = null;
        for(int i = 0; i < 15; i++)
        {
            try
            {
                LOGGER.info("Retrieving books from returned supplier");
                books = booksSupplier.get();
            }
            catch(Exception e)
            {
                LOGGER.error("Could not retrieve books from supplier", e);
            }
        }
        model.addAttribute("books", books);

        return "books";
    }
}

So when the user clicks on the books page, we retrieve books from our BooksApplication REST Service.

I have autowired the bean for countCircuitBreaker. For demo purposes – I will be calling the REST service 15 times in a loop to get all the books. This way, I can simulate interruption on my REST service side.

Our circuit breaker decorates a supplier that does REST call to remote service and the supplier stores the result of our remote service call.

In this demo, we are calling our REST service in a sequential manner, but remote service calls can happen parallelly also. The circuit breaker will still keep track of results irrespective of sequential or parallel calls.

Demo

Let’s look at how the circuit breaker will function in a live demo now. My REST service is running on port 8443 and my Circuitbreakerdemo application is running on port 8743.

Initially, I start both of the applications and access the home page of Circuitbreakerdemo application. The home page contains the link for viewing all the books from the store.

Now to simulate some errors, I have added the following code in my RestTemplate call that basically sleeps for 3 seconds before returning the result of the REST call.

    public List getAllBooksFromLibrary ()
    {
        HttpHeaders httpHeaders = new HttpHeaders();
        httpHeaders.setContentType(MediaType.APPLICATION_JSON);

        ResponseEntity<List> responseEntity;
        long startTime = System.currentTimeMillis();
        LOGGER.info("Start time = {}", startTime);
        try
        {
            responseEntity= restTemplate.exchange(buildUrl(),
                    HttpMethod.GET, null, new ParameterizedTypeReference<List>()
                    {});
            if(responseEntity != null && responseEntity.hasBody())
            {
                Thread.sleep(3000);
                LOGGER.info("Total time to retrieve results = {}",
                        System.currentTimeMillis() - startTime);
                return responseEntity.getBody();
            }
        }
        catch (URISyntaxException | InterruptedException e)
        {
            LOGGER.error("URI has a wrong syntax", e);
        }

        LOGGER.info("No result found, returning an empty list");
        return new ArrayList<>();
    }

In short, my circuit breaker loop will call the service enough times to pass the threshold of 65 percent of slow calls that are of duration more than 3 seconds. Once I click on the link for here, I will receive the result, but my circuit breaker will be open and will not allow future calls till it is in either half-open state or closed state.

You will notice that we started getting an exception CallNotPermittedException when the circuit breaker was in the OPEN state. Also, the circuit breaker was opened when the 10 calls were performed. This is because our sliding window size is 10.

Another way, I can simulate the error by shutting down my REST service or database service. That way REST calls can take longer than required.

Now, let’s switch the COUNT_BASED circuit breaker to TIME_BASED circuit breaker. In TIME_BASED circuit breaker, we will switch off our REST service after a second, and then we will click on here link from the home page. If 70 percent of calls in the last 10 seconds fail, our circuit breaker will open.

Since REST Service is closed, we will see the following errors in Circuitbreakdemo application

We will see the number of errors before the circuit breaker will be in OPEN state.

One configuration we can always add how long we want to keep the circuit breaker in the open state.  For the demo, I have added the circuit breaker will be in an open state for 10 seconds.

How to handle OPEN circuit breakers?

One question arises, how do you handle OPEN circuit breakers? Luckily, resilience4j offers a fallback configuration with Decorators utility. In most cases, you can always configure this to get the result from previous successful results so that users can still work with the application.

Conclusion

In this post, I have covered how to use a circuit breaker in a Spring Boot application. The code for this demo is available here.

In this demo, I have not covered how to monitor these circuit breaker events as resilience4j the library allows storing these events with metrics that one can monitor with a monitoring system.

If you enjoyed this post, consider subscribing to my blog here.

References

1. Resilience4J Library – Resilience4J
2. Circuit Breaker with Resilience4j – Circuit Breaker

In this post, I want to show how you can use Spring Batch. This is a step by step Spring Batch Tutorial.

In enterprise applications, batch processing is common. But with data becoming more prevalent on the internet, it has also become important how we process this data. There are multiple solutions available. Apache Storm or Apache Spark helps with processing and transforming the data in the required format. In this post, we will be looking at Spring Batch more closely.

What is Spring Batch?

Spring Batch is a lightweight framework designed to facilitate batch processing. It allows developers to create batch applications. In turn, these batch applications process the incoming data and transform it for further usage.

Another big advantage of using the Spring Batch is that it allows for high-performance processing of this data. The applications that rely upon data heavily, it is of utmost importance that data becomes instantly available.

Spring Batch allows a developer to use POJO based approach. In this approach, a developer can transform the batch-processed data into data models that she can further use for application business logic.

In this post, I will cover an example where we will batch process a data-intensive CSV file for employee records and transform, and validate that data to load into our database.

What is Batch Processing?

Batch processing is a data processing mode. It involves consuming all the data, processing that data, transforming it, and then sending it to another data source. Usually, this is done through an automated job. Either a triggering system or a user triggers a job and that job processes the job definition. Job definition will be about consuming the data from its source.

The key advantage of batch processing is it handles a large volume of data. Nevertheless, this operation can be asynchronous. Most applications perform batch processing separately from real-time user interaction.

Next, we will learn about the Spring Batch framework and what it comprises.

Spring Batch Framework

The following architecture shows the components of the Spring Batch framework.

First, the batch process involves a job. User schedules a job to be run at a certain time or based on a certain condition. This can also involve a job trigger.

Spring Batch framework also includes

• logging and tracing
• transaction management
• job processing statistics
• job restart
• resource management

Usually, when you configure a job, it will be saved in the job repository. Job Repository keeps the metadata information of all the jobs. A trigger starts these jobs at their scheduled time.

A job launcher is an interface to launch a job or runs a job when the jobs’ scheduled time arrives.

Job is defined with job parameters. When a job starts, a job instance runs for that job. Every execution of job instance has job execution and it keeps track status of the job. A job can have multiple steps.

Step is an independent phase of a job. A job can be comprised of more than one step. Similar to the job, each step has step execution that executes the step and keeps track of the status of the step.

Each step has an item reader that basically reads the input data, an item processor that processes the data and transforms it, and an item writer that takes the processed data and output it.

Now, let’s see all these components in our demo.

Step by Step Spring Batch Tutorial with an example

As part of the demo, we will be uploading a csv file through Spring Batch Framework. So to start with, create the spring project and add the following dependency:

implementation 'org.springframework.boot:spring-boot-starter-batch'

This is the main dependency of our project. Also our main application will look like below:

package com.betterjavacode.springbatchdemo;

import org.springframework.boot.SpringApplication;
import org.springframework.boot.autoconfigure.SpringBootApplication;


@SpringBootApplication
public class SpringbatchdemoApplication
{

 public static void main(String[] args)
 {
  SpringApplication.run(SpringbatchdemoApplication.class, args);
 }

}

Create DTO Object

I will be uploading employee data through a CSV file, so I will have my DTO object for Employee created as below:

package com.betterjavacode.springbatchdemo.dtos;

import com.betterjavacode.springbatchdemo.models.Company;
import com.betterjavacode.springbatchdemo.models.Employee;
import com.betterjavacode.springbatchdemo.repositories.CompanyRepository;
import org.springframework.beans.factory.annotation.Autowired;


import java.io.Serializable;

public class EmployeeDto implements Serializable
{
    private static final long serialVersionUID = 710566148641281929L;

    @Autowired
    public CompanyRepository companyRepository;

    private int employeeId;
    private int companyId;
    private String firstName;
    private String lastName;
    private String email;
    private String jobTitle;

    public EmployeeDto()
    {

    }

    public EmployeeDto(int employeeId, String firstName, String lastName, String email,
                        String jobTitle, int companyId)
    {
        this.employeeId = employeeId;
        this.firstName = firstName;
        this.lastName = lastName;
        this.email = email;
        this.jobTitle = jobTitle;
        this.companyId = companyId;
    }

    public Employee employeeDtoToEmployee()
    {
        Employee employee = new Employee();
        employee.setEmployeeId(this.employeeId);
        employee.setFirstName(this.firstName);
        employee.setLastName(this.lastName);
        employee.setEmail(this.email);
        Company company = companyRepository.findById(this.companyId).get();
        employee.setCompany(company);
        employee.setJobTitle(this.jobTitle);
        return employee;
    }

    public int getEmployeeId ()
    {
        return employeeId;
    }

    public void setEmployeeId (int employeeId)
    {
        this.employeeId = employeeId;
    }

    public int getCompanyId ()
    {
        return companyId;
    }

    public void setCompanyId (int companyId)
    {
        this.companyId = companyId;
    }

    public String getFirstName ()
    {
        return firstName;
    }

    public void setFirstName (String firstName)
    {
        this.firstName = firstName;
    }

    public String getLastName ()
    {
        return lastName;
    }

    public void setLastName (String lastName)
    {
        this.lastName = lastName;
    }

    public String getEmail ()
    {
        return email;
    }

    public void setEmail (String email)
    {
        this.email = email;
    }

    public String getJobTitle ()
    {
        return jobTitle;
    }

    public void setJobTitle (String jobTitle)
    {
        this.jobTitle = jobTitle;
    }
}

This DTO class also uses a repository CompanyRepository to get a company object and convert DTO to a database object.

Setting up Spring Batch Configuration

Now, we will set up a batch configuration for our job that will run to upload a CSV file into the database. Our class BatchConfig contain an annotation @EnableBatchProcessing. This annotation enables Spring Batch features and provides a base configuration to set up batch jobs in a @Configuration class.

@Configuration
@EnableBatchProcessing
public class BatchConfig
{

}

This Batch Configuration will include a definition of our job, steps involved in the job. It will also include how we want to read our file data and process it further.

    @Bean
    public Job processJob(Step step)
    {
        return jobBuilderFactory.get("processJob")
                .incrementer(new RunIdIncrementer())
                .listener(listener())
                .flow(step).end().build();
    }

    @Bean
    public Step orderStep1(JdbcBatchItemWriter writer)
    {
        return stepBuilderFactory.get("orderStep1").<EmployeeDto, EmployeeDto> chunk(10)
                .reader(flatFileItemReader())
                .processor(employeeItemProcessor())
                .writer(writer).build();
    }

Above bean declares the job processJob. incrementer adds job parameters. listener will listen to job and handle job status. The bean for listener will handle job completion or job failure notification. As discussed in Spring Batch architecture, every job includes more than one step.

@Bean for step uses stepBuilderFactory to create a step. This step processes a chunk of data in a size of 10. It has a Flat File Reader flatFileItemReader(). A processor employeeItemReader will process the data that has been read by Flat File Item Reader.

    @Bean
    public FlatFileItemReader flatFileItemReader()
    {
        return new FlatFileItemReaderBuilder()
                .name("flatFileItemReader")
                .resource(new ClassPathResource("input/employeedata.csv"))
                .delimited()
                .names(format)
                .linesToSkip(1)
                .lineMapper(lineMapper())
                .fieldSetMapper(new BeanWrapperFieldSetMapper(){{
                    setTargetType(EmployeeDto.class);
                }})
                .build();
    }

    @Bean
    public LineMapper lineMapper()
    {
        final DefaultLineMapper defaultLineMapper = new DefaultLineMapper<>();
        final DelimitedLineTokenizer delimitedLineTokenizer = new DelimitedLineTokenizer();
        delimitedLineTokenizer.setDelimiter(",");
        delimitedLineTokenizer.setStrict(false);
        delimitedLineTokenizer.setNames(format);

        defaultLineMapper.setLineTokenizer(delimitedLineTokenizer);
        defaultLineMapper.setFieldSetMapper(employeeDtoFieldSetMapper);

        return defaultLineMapper;
    }

    @Bean
    public EmployeeItemProcessor employeeItemProcessor()
    {
        return new EmployeeItemProcessor();
    }

    @Bean
    public JobExecutionListener listener()
    {
        return new JobCompletionListener();
    }

    @Bean
    public JdbcBatchItemWriter writer(final DataSource dataSource)
    {
        return new JdbcBatchItemWriterBuilder()
                .itemSqlParameterSourceProvider(new BeanPropertyItemSqlParameterSourceProvider<>())
                .sql("INSERT INTO employee(employeeId, firstName, lastName, jobTitle, email, " +
                        "companyId) VALUES(:employeeId, :firstName, :lastName, :jobTitle, :email," +
                        " " +
                        ":companyId)")
                .dataSource(dataSource)
                .build();
    }

We will take a look at each of these beans now.

FlatFileItemReader will read the data from the flat file. We are using a FlatFileItemReaderBuilder to create a FlatFileItemReader of type EmployeeDto.

resource indicates the location of the file.

delimited – This builds a delimited tokenizer.

names – will show the order of fields in the file.

lineMapper is an interface to map lines from file to domain object.

fieldSetMapper will map the data from fieldset to an object.

lineMapper bean needs tokenizer and fieldsetmapper.

employeeDtoFieldSetMapper is another bean that we have autowired in this class.

package com.betterjavacode.springbatchdemo.configurations.processor;

import com.betterjavacode.springbatchdemo.dtos.EmployeeDto;
import org.springframework.batch.item.file.mapping.FieldSetMapper;
import org.springframework.batch.item.file.transform.FieldSet;
import org.springframework.stereotype.Component;
import org.springframework.validation.BindException;

@Component
public class EmployeeDtoFieldSetMapper implements FieldSetMapper
{

    @Override
    public EmployeeDto mapFieldSet (FieldSet fieldSet) throws BindException
    {
        int employeeId = fieldSet.readInt("employeeId");
        String firstName = fieldSet.readRawString("firstName");
        String lastName = fieldSet.readRawString("lastName");
        String jobTitle = fieldSet.readRawString("jobTitle");
        String email = fieldSet.readRawString("email");
        int companyId = fieldSet.readInt("companyId");

        return new EmployeeDto(employeeId, firstName, lastName, jobTitle, email, companyId);
    }
}

As you can see, this FieldSetMapper maps fields to individual objects to create an EmployeeDto.

EmployeeItemProcessor implements the interface ItemProcessor. Basically in this class, we validate EmployeeDto data to verify if the company, the employee belongs to, exists.

JobCompletionListener  checks for job completion status.

    @Override
    public void afterJob(JobExecution jobExecution)
    {
        if (jobExecution.getStatus() == BatchStatus.COMPLETED)
        {
            // Log statement
            System.out.println("BATCH JOB COMPLETED SUCCESSFULLY");
        }
    }

Now, let’s look at ItemWriter. This bean basically uses JdbcBatchItemWriter  . JdbcBatchItemWriter uses INSERT sql statement to insert processed EmployeeDto data into the configured data source.

Configuring Application Properties

Before we run our application to process a file, let’s look at application.properties.

spring.datasource.url=jdbc:mysql://127.0.0.1/springbatchdemo?autoReconnect=true&useSSL=false
spring.datasource.username = root
spring.datasource.password=*******
spring.datasource.driver-class-name=com.mysql.cj.jdbc.Driver
spring.jpa.show-sql=true
spring.jpa.properties.hibernate.dialect = org.hibernate.dialect.MySQL5Dialect
spring.datasource.hikari.connection-test-query=SELECT 1
spring.batch.initialize-schema=ALWAYS

Other than regular data source properties, we should understand the property spring.batch.initialize-schema=ALWAYS.  If we don’t use this property and start the application, the application will complain Table batch_job_instance doesn't exist.

To avoid this error, we are basically telling to create batch job-related metadata during startup. This property will create additional database tables in your database like batch_job_execution, batch_job_execution_context, batch_job_execution_params, batch_job_instance etc.

Demo

Now if I execute my Spring Boot Application, it will run and execute the job. There are different ways to trigger a job. In an enterprise application, you will receive a file or data in some kind of storage place (S3 or Amazon SNS-SQS), and you will have a job that will be monitoring this location to trigger the file loading Spring Batch job.

You can see in the execution a message about job completion – “BATCH JOB COMPLETED SUCCESSFULLY“. If we check our database table, we will see the data loaded.

You can download the code for this demo from my github repository.

What more?

I have covered a Spring Batch tutorial here, but this is not all. There is more to Spring Batch than this introductory part. You can have different input data sources or you can also load the data from file to file with various data processing rules.

There are also ways to automate these jobs and process a high volume of data in a performant manner.

Conclusion

In this post, I showed a step by step Spring Batch Tutorial. There are many ways to handle batch jobs, but Spring Batch has made this very easy.

In other news, I recently released my new book – Simplifying Spring Security. If you are looking to learn about Spring Security, you can buy the book here. Accompany this book with this post of  Spring Boot Interview questions and you will be ready for your next job interview.