There are a number of pre-built implementations of TaskExecutor included with the Spring distribution. In all likelihood, you should never need to implement your own. The common out-of-the-box variants are:

Spring’s TaskExecutor implementations are used as simple JavaBeans. In the example below, we define a bean that uses the ThreadPoolTaskExecutor to asynchronously print out a set of messages.

import org.springframework.core.task.TaskExecutor;

public class TaskExecutorExample {

    private class MessagePrinterTask implements Runnable {

        private String message;

        public MessagePrinterTask(String message) {
            this.message = message;
        }

        public void run() {
            System.out.println(message);
        }
    }

    private TaskExecutor taskExecutor;

    public TaskExecutorExample(TaskExecutor taskExecutor) {
        this.taskExecutor = taskExecutor;
    }

    public void printMessages() {
        for(int i = 0; i < 25; i++) {
            taskExecutor.execute(new MessagePrinterTask("Message" + i));
        }
    }
}

As you can see, rather than retrieving a thread from the pool and executing yourself, you add your Runnable to the queue and the TaskExecutor uses its internal rules to decide when the task gets executed.

To configure the rules that the TaskExecutor will use, simple bean properties have been exposed.

<bean id="taskExecutor" class="org.springframework.scheduling.concurrent.ThreadPoolTaskExecutor">
    <property name="corePoolSize" value="5"/>
    <property name="maxPoolSize" value="10"/>
    <property name="queueCapacity" value="25"/>
</bean>

<bean id="taskExecutorExample" class="TaskExecutorExample">
    <constructor-arg ref="taskExecutor"/>
</bean>

public interface TaskScheduler {

    ScheduledFuture schedule(Runnable task, Trigger trigger);

    ScheduledFuture schedule(Runnable task, Date startTime);

    ScheduledFuture scheduleAtFixedRate(Runnable task, Date startTime, long period);

    ScheduledFuture scheduleAtFixedRate(Runnable task, long period);

    ScheduledFuture scheduleWithFixedDelay(Runnable task, Date startTime, long delay);

    ScheduledFuture scheduleWithFixedDelay(Runnable task, long delay);
}

The Trigger interface is essentially inspired by JSR-236 which, as of Spring 3.0, was not yet officially implemented. The basic idea of the Trigger is that execution times may be determined based on past execution outcomes or even arbitrary conditions. If these determinations do take into account the outcome of the preceding execution, that information is available within a TriggerContext. The Trigger interface itself is quite simple:

public interface Trigger {

    Date nextExecutionTime(TriggerContext triggerContext);
}

As you can see, the TriggerContext is the most important part. It encapsulates all of the relevant data, and is open for extension in the future if necessary. The TriggerContext is an interface (a SimpleTriggerContext implementation is used by default). Here you can see what methods are available for Trigger implementations.

public interface TriggerContext {

    Date lastScheduledExecutionTime();

    Date lastActualExecutionTime();

    Date lastCompletionTime();
}

Spring provides two implementations of the Trigger interface. The most interesting one is the CronTrigger. It enables the scheduling of tasks based on cron expressions. For example, the following task is being scheduled to run 15 minutes past each hour but only during the 9-to-5 "business hours" on weekdays.

scheduler.schedule(task, new CronTrigger("0 15 9-17 * * MON-FRI"));

The other out-of-the-box implementation is a PeriodicTrigger that accepts a fixed period, an optional initial delay value, and a boolean to indicate whether the period should be interpreted as a fixed-rate or a fixed-delay. Since the TaskScheduler interface already defines methods for scheduling tasks at a fixed-rate or with a fixed-delay, those methods should be used directly whenever possible. The value of the PeriodicTrigger implementation is that it can be used within components that rely on the Trigger abstraction. For example, it may be convenient to allow periodic triggers, cron-based triggers, and even custom trigger implementations to be used interchangeably. Such a component could take advantage of dependency injection so that such Triggers could be configured externally and therefore easily modified or extended.

As with Spring’s TaskExecutor abstraction, the primary benefit of the TaskScheduler arrangement is that an application’s scheduling needs are decoupled from the deployment environment. This abstraction level is particularly relevant when deploying to an application server environment where threads should not be created directly by the application itself. For such scenarios, Spring provides a TimerManagerTaskScheduler delegating to a CommonJ TimerManager on WebLogic/WebSphere as well as a more recent DefaultManagedTaskScheduler delegating to a JSR-236 ManagedScheduledExecutorService in a Java EE 7+ environment, both typically configured with a JNDI lookup.

Whenever external thread management is not a requirement, a simpler alternative is a local ScheduledExecutorService setup within the application which can be adapted through Spring’s ConcurrentTaskScheduler. As a convenience, Spring also provides a ThreadPoolTaskScheduler which internally delegates to a ScheduledExecutorService, providing common bean-style configuration along the lines of ThreadPoolTaskExecutor. These variants work perfectly fine for locally embedded thread pool setups in lenient application server environments as well, in particular on Tomcat and Jetty.

To enable support for @Scheduled and @Async annotations add @EnableScheduling and @EnableAsync to one of your @Configuration classes:

@Configuration
@EnableAsync
@EnableScheduling
public class AppConfig {
}

You are free to pick and choose the relevant annotations for your application. For example, if you only need support for @Scheduled, simply omit @EnableAsync. For more fine-grained control you can additionally implement the SchedulingConfigurer and/or AsyncConfigurer interfaces. See the javadocs for full details.

If you prefer XML configuration use the <task:annotation-driven> element.

<task:annotation-driven executor="myExecutor" scheduler="myScheduler"/>
<task:executor id="myExecutor" pool-size="5"/>
<task:scheduler id="myScheduler" pool-size="10"/>

Notice with the above XML that an executor reference is provided for handling those tasks that correspond to methods with the @Async annotation, and the scheduler reference is provided for managing those methods annotated with @Scheduled.

The @Scheduled annotation can be added to a method along with trigger metadata. For example, the following method would be invoked every 5 seconds with a fixed delay, meaning that the period will be measured from the completion time of each preceding invocation.

@Scheduled(fixedDelay=5000)
public void doSomething() {
    // something that should execute periodically
}

If a fixed rate execution is desired, simply change the property name specified within the annotation. The following would be executed every 5 seconds measured between the successive start times of each invocation.

@Scheduled(fixedRate=5000)
public void doSomething() {
    // something that should execute periodically
}

For fixed-delay and fixed-rate tasks, an initial delay may be specified indicating the number of milliseconds to wait before the first execution of the method.

@Scheduled(initialDelay=1000, fixedRate=5000)
public void doSomething() {
    // something that should execute periodically
}

If simple periodic scheduling is not expressive enough, then a cron expression may be provided. For example, the following will only execute on weekdays.

@Scheduled(cron="*/5 * * * * MON-FRI")
public void doSomething() {
    // something that should execute on weekdays only
}

Notice that the methods to be scheduled must have void returns and must not expect any arguments. If the method needs to interact with other objects from the Application Context, then those would typically have been provided through dependency injection.

The @Async annotation can be provided on a method so that invocation of that method will occur asynchronously. In other words, the caller will return immediately upon invocation and the actual execution of the method will occur in a task that has been submitted to a Spring TaskExecutor. In the simplest case, the annotation may be applied to a void-returning method.

@Async
void doSomething() {
    // this will be executed asynchronously
}

Unlike the methods annotated with the @Scheduled annotation, these methods can expect arguments, because they will be invoked in the "normal" way by callers at runtime rather than from a scheduled task being managed by the container. For example, the following is a legitimate application of the @Async annotation.

@Async
void doSomething(String s) {
    // this will be executed asynchronously
}

Even methods that return a value can be invoked asynchronously. However, such methods are required to have a Future typed return value. This still provides the benefit of asynchronous execution so that the caller can perform other tasks prior to calling get() on that Future.

@Async
Future<String> returnSomething(int i) {
    // this will be executed asynchronously
}

@Async can not be used in conjunction with lifecycle callbacks such as @PostConstruct. To asynchronously initialize Spring beans you currently have to use a separate initializing Spring bean that invokes the @Async annotated method on the target then.

public class SampleBeanImpl implements SampleBean {

    @Async
    void doSomething() {
        // ...
    }

}

public class SampleBeanInitializer {

    private final SampleBean bean;

    public SampleBeanInitializer(SampleBean bean) {
        this.bean = bean;
    }

    @PostConstruct
    public void initialize() {
        bean.doSomething();
    }

}

By default when specifying @Async on a method, the executor that will be used is the one supplied to the 'annotation-driven' element as described above. However, the value attribute of the @Async annotation can be used when needing to indicate that an executor other than the default should be used when executing a given method.

@Async("otherExecutor")
void doSomething(String s) {
    // this will be executed asynchronously by "otherExecutor"
}

In this case, "otherExecutor" may be the name of any Executor bean in the Spring container, or may be the name of a qualifier associated with any Executor, e.g. as specified with the <qualifier> element or Spring’s @Qualifier annotation.

When an @Async method has a Future typed return value, it is easy to manage an exception that was thrown during the method execution as this exception will be thrown when calling get on the Future result. With a void return type however, the exception is uncaught and cannot be transmitted. For those cases, an AsyncUncaughtExceptionHandler can be provided to handle such exceptions.

public class MyAsyncUncaughtExceptionHandler implements AsyncUncaughtExceptionHandler {

    @Override
    public void handleUncaughtException(Throwable ex, Method method, Object... params) {
        // handle exception
    }
}

By default, the exception is simply logged. A custom AsyncUncaughtExceptionHandler can be defined via AsyncConfigurer or the task:annotation-driven XML element.

The following element will create a ThreadPoolTaskScheduler instance with the specified thread pool size.

<task:scheduler id="scheduler" pool-size="10"/>

The value provided for the 'id' attribute will be used as the prefix for thread names within the pool. The 'scheduler' element is relatively straightforward. If you do not provide a 'pool-size' attribute, the default thread pool will only have a single thread. There are no other configuration options for the scheduler.

The following will create a ThreadPoolTaskExecutor instance:

<task:executor id="executor" pool-size="10"/>

As with the scheduler above, the value provided for the 'id' attribute will be used as the prefix for thread names within the pool. As far as the pool size is concerned, the 'executor' element supports more configuration options than the 'scheduler' element. For one thing, the thread pool for a ThreadPoolTaskExecutor is itself more configurable. Rather than just a single size, an executor’s thread pool may have different values for the core and the max size. If a single value is provided then the executor will have a fixed-size thread pool (the core and max sizes are the same). However, the 'executor' element’s 'pool-size' attribute also accepts a range in the form of "min-max".

<task:executor
        id="executorWithPoolSizeRange"
        pool-size="5-25"
        queue-capacity="100"/>

As you can see from that configuration, a 'queue-capacity' value has also been provided. The configuration of the thread pool should also be considered in light of the executor’s queue capacity. For the full description of the relationship between pool size and queue capacity, consult the documentation for ThreadPoolExecutor. The main idea is that when a task is submitted, the executor will first try to use a free thread if the number of active threads is currently less than the core size. If the core size has been reached, then the task will be added to the queue as long as its capacity has not yet been reached. Only then, if the queue’s capacity has been reached, will the executor create a new thread beyond the core size. If the max size has also been reached, then the executor will reject the task.

By default, the queue is unbounded, but this is rarely the desired configuration, because it can lead to OutOfMemoryErrors if enough tasks are added to that queue while all pool threads are busy. Furthermore, if the queue is unbounded, then the max size has no effect at all. Since the executor will always try the queue before creating a new thread beyond the core size, a queue must have a finite capacity for the thread pool to grow beyond the core size (this is why a fixed size pool is the only sensible case when using an unbounded queue).

In a moment, we will review the effects of the keep-alive setting which adds yet another factor to consider when providing a pool size configuration. First, let’s consider the case, as mentioned above, when a task is rejected. By default, when a task is rejected, a thread pool executor will throw a TaskRejectedException. However, the rejection policy is actually configurable. The exception is thrown when using the default rejection policy which is the AbortPolicy implementation. For applications where some tasks can be skipped under heavy load, either the DiscardPolicy or DiscardOldestPolicy may be configured instead. Another option that works well for applications that need to throttle the submitted tasks under heavy load is the CallerRunsPolicy. Instead of throwing an exception or discarding tasks, that policy will simply force the thread that is calling the submit method to run the task itself. The idea is that such a caller will be busy while running that task and not able to submit other tasks immediately. Therefore it provides a simple way to throttle the incoming load while maintaining the limits of the thread pool and queue. Typically this allows the executor to "catch up" on the tasks it is handling and thereby frees up some capacity on the queue, in the pool, or both. Any of these options can be chosen from an enumeration of values available for the 'rejection-policy' attribute on the 'executor' element.

<task:executor
        id="executorWithCallerRunsPolicy"
        pool-size="5-25"
        queue-capacity="100"
        rejection-policy="CALLER_RUNS"/>

Finally, the keep-alive setting determines the time limit (in seconds) for which threads may remain idle before being terminated. If there are more than the core number of threads currently in the pool, after waiting this amount of time without processing a task, excess threads will get terminated. A time value of zero will cause excess threads to terminate immediately after executing a task without remaining follow-up work in the task queue.

<task:executor
        id="executorWithKeepAlive"
        pool-size="5-25"
        keep-alive="120"/>

The most powerful feature of Spring’s task namespace is the support for configuring tasks to be scheduled within a Spring Application Context. This follows an approach similar to other "method-invokers" in Spring, such as that provided by the JMS namespace for configuring Message-driven POJOs. Basically a "ref" attribute can point to any Spring-managed object, and the "method" attribute provides the name of a method to be invoked on that object. Here is a simple example.

<task:scheduled-tasks scheduler="myScheduler">
    <task:scheduled ref="beanA" method="methodA" fixed-delay="5000"/>
</task:scheduled-tasks>

<task:scheduler id="myScheduler" pool-size="10"/>

As you can see, the scheduler is referenced by the outer element, and each individual task includes the configuration of its trigger metadata. In the preceding example, that metadata defines a periodic trigger with a fixed delay indicating the number of milliseconds to wait after each task execution has completed. Another option is 'fixed-rate', indicating how often the method should be executed regardless of how long any previous execution takes. Additionally, for both fixed-delay and fixed-rate tasks an 'initial-delay' parameter may be specified indicating the number of milliseconds to wait before the first execution of the method. For more control, a "cron" attribute may be provided instead. Here is an example demonstrating these other options.

<task:scheduled-tasks scheduler="myScheduler">
    <task:scheduled ref="beanA" method="methodA" fixed-delay="5000" initial-delay="1000"/>
    <task:scheduled ref="beanB" method="methodB" fixed-rate="5000"/>
    <task:scheduled ref="beanC" method="methodC" cron="*/5 * * * * MON-FRI"/>
</task:scheduled-tasks>

<task:scheduler id="myScheduler" pool-size="10"/>

Quartz JobDetail objects contain all information needed to run a job. Spring provides a JobDetailFactoryBean which provides bean-style properties for XML configuration purposes. Let’s have a look at an example:

<bean name="exampleJob" class="org.springframework.scheduling.quartz.JobDetailFactoryBean">
    <property name="jobClass" value="example.ExampleJob"/>
    <property name="jobDataAsMap">
        <map>
            <entry key="timeout" value="5"/>
        </map>
    </property>
</bean>

The job detail configuration has all information it needs to run the job (ExampleJob). The timeout is specified in the job data map. The job data map is available through the JobExecutionContext (passed to you at execution time), but the JobDetail also gets its properties from the job data mapped to properties of the job instance. So in this case, if the ExampleJob contains a bean property named timeout, the JobDetail will have it applied automatically:

package example;

public class ExampleJob extends QuartzJobBean {

    private int timeout;

    /**
     * Setter called after the ExampleJob is instantiated
     * with the value from the JobDetailFactoryBean (5)
     */
    public void setTimeout(int timeout) {
        this.timeout = timeout;
    }

    protected void executeInternal(JobExecutionContext ctx) throws JobExecutionException {
        // do the actual work
    }

}

All additional properties from the job data map are of course available to you as well.

Often you just need to invoke a method on a specific object. Using the MethodInvokingJobDetailFactoryBean you can do exactly this:

<bean id="jobDetail" class="org.springframework.scheduling.quartz.MethodInvokingJobDetailFactoryBean">
    <property name="targetObject" ref="exampleBusinessObject"/>
    <property name="targetMethod" value="doIt"/>
</bean>

The above example will result in the doIt method being called on the exampleBusinessObject method (see below):

public class ExampleBusinessObject {

    // properties and collaborators

    public void doIt() {
        // do the actual work
    }
}

<bean id="exampleBusinessObject" class="examples.ExampleBusinessObject"/>

Using the MethodInvokingJobDetailFactoryBean, you don’t need to create one-line jobs that just invoke a method, and you only need to create the actual business object and wire up the detail object.

By default, Quartz Jobs are stateless, resulting in the possibility of jobs interfering with each other. If you specify two triggers for the same JobDetail, it might be possible that before the first job has finished, the second one will start. If JobDetail classes implement the Stateful interface, this won’t happen. The second job will not start before the first one has finished. To make jobs resulting from the MethodInvokingJobDetailFactoryBean non-concurrent, set the concurrent flag to false.

<bean id="jobDetail" class="org.springframework.scheduling.quartz.MethodInvokingJobDetailFactoryBean">
    <property name="targetObject" ref="exampleBusinessObject"/>
    <property name="targetMethod" value="doIt"/>
    <property name="concurrent" value="false"/>
</bean>

We’ve created job details and jobs. We’ve also reviewed the convenience bean that allows you to invoke a method on a specific object. Of course, we still need to schedule the jobs themselves. This is done using triggers and a SchedulerFactoryBean. Several triggers are available within Quartz and Spring offers two Quartz FactoryBean implementations with convenient defaults: CronTriggerFactoryBean and SimpleTriggerFactoryBean.

Triggers need to be scheduled. Spring offers a SchedulerFactoryBean that exposes triggers to be set as properties. SchedulerFactoryBean schedules the actual jobs with those triggers.

Find below a couple of examples:

<bean id="simpleTrigger" class="org.springframework.scheduling.quartz.SimpleTriggerFactoryBean">
    <!-- see the example of method invoking job above -->
    <property name="jobDetail" ref="jobDetail"/>
    <!-- 10 seconds -->
    <property name="startDelay" value="10000"/>
    <!-- repeat every 50 seconds -->
    <property name="repeatInterval" value="50000"/>
</bean>

<bean id="cronTrigger" class="org.springframework.scheduling.quartz.CronTriggerFactoryBean">
    <property name="jobDetail" ref="exampleJob"/>
    <!-- run every morning at 6 AM -->
    <property name="cronExpression" value="0 0 6 * * ?"/>
</bean>

Now we’ve set up two triggers, one running every 50 seconds with a starting delay of 10 seconds and one every morning at 6 AM. To finalize everything, we need to set up the SchedulerFactoryBean:

<bean class="org.springframework.scheduling.quartz.SchedulerFactoryBean">
    <property name="triggers">
        <list>
            <ref bean="cronTrigger"/>
            <ref bean="simpleTrigger"/>
        </list>
    </property>
</bean>

More properties are available for the SchedulerFactoryBean for you to set, such as the calendars used by the job details, properties to customize Quartz with, etc. Have a look at the SchedulerFactoryBean javadocs for more information.