Performance issues in SimpleJobRepository.getStepExecutionCount(JobInstance jobInstance, String stepName)

[bcubk]

Hi Spring-Batch Team,

couple days ago, we noticed a high amount of CPU utilization in our Spring Batch application. Our batch job contains a high amount of job steps. We debugged the code and localized a small issue in class SimpleJobRepository.getStepExecutionCount(JobInstance jobInstance, String stepName). Here the implementation:

/**
* @return number of executions of the step within given job instance
*/
@Override
public int getStepExecutionCount(JobInstance jobInstance, String stepName) {
	int count = 0;
	List<JobExecution> jobExecutions = jobExecutionDao.findJobExecutions(jobInstance);
	for (JobExecution jobExecution : jobExecutions) {
		stepExecutionDao.addStepExecutions(jobExecution);
		for (StepExecution stepExecution : jobExecution.getStepExecutions()) {
			if (stepName.equals(stepExecution.getStepName())) {
				count++;
			}
		}
	}
	return count;
}

This retrieves all StepExecutions for all JobExecutions of a particular JobInstance and counts the StepExecutions in a for-loop. We changed this method and executing a count in the underlying database which solved our CPU problem. Is this a good solution? If not, why? Of course, we would contribute our fix 👍

Thanks in advance and kind regards
Baris

[fmbenhassine]

Hi @bcubk ,

Thank you for opening this issue. Indeed, this method can be optimized by using a single count query to the database. You are welcome to contribute if you want!

Best regards,
Mahmoud