com.mulesoft.anypoint.retry

Class BackoffRunnableRetrier<T>

java.lang.Object

com.mulesoft.anypoint.retry.BackoffRunnableRetrier<T>

All Implemented Interfaces:

FastRecoveryObserver, RunnableRetrier<T>, org.mule.runtime.api.lifecycle.Disposable

public class BackoffRunnableRetrier<T>
extends Object
implements RunnableRetrier<T>, FastRecoveryObserver

Executes multiple Runnables in a separate thread until they complete successfully. An execution is considered successful when it does not end with an Exception.

Under the hood, this object uses the BackoffScheduler's engine, in FastRecovery mode. This means that backoff will be applied to those task who raise an Exception. The first execution that finishes without an error, will mean that it ended successfully and will not be rescheduled again. BackoffConfiguration can be set to each instance of this.

BackoffRunnableRetrier handles one queue per instance of key (of type T). Subsequent queued Runnables for a key will not be executed until its current scheduled task reaches the Stable state. NOTE: if for a certain key, running task never halts, the remaining queued tasks will starve.

To minimize runtime resource consumption, BackoffScheduler will be disposed if there are no further tasks to process. It will be recreated as soon as a Runnable is scheduled.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	BackoffRunnableRetrier.Builder<T>

Field Summary

Fields inherited from interface org.mule.runtime.api.lifecycle.Disposable

PHASE_NAME

Method Summary

Modifier and Type	Method and Description
static SchedulingConfiguration	delayInitialScheduling(int initialDelay)
void	dispose()
FastRecoveryObserver	fastRecoveryAbort(BackoffRunnable backoffRunnable) Notifies the fast recovery BackoffRunnable is now with Error.
FastRecoveryObserver	fastRecoveryStable(BackoffRunnable backoffRunnable, FastRecoveryConfiguration configuration) Notifies the fast recovery BackoffRunnable is now Stable.
FastRecoveryObserver	fastRecoveryUnstable(BackoffRunnable runnable, FastRecoveryConfiguration configuration) Notifies the fast recovery BackoffRunnable is still Unstable.
boolean	hasQueuedRunnables(T key) Return whether there are task pending to execute for an specific key.
boolean	isIdle(T key) Return whether the retrier is idle or there are tasks running or pending for execution for a specific key.
RunnableRetrier<T>	scheduleRetry(T key, Runnable runnable) Schedules a task to be executed until successful.
static SchedulingConfiguration	zeroDelayOnScheduling()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Detail

scheduleRetry

public RunnableRetrier<T> scheduleRetry(T key,
                                        Runnable runnable)

Schedules a task to be executed until successful. If there is no Runnable already running for that key, task will be immediately scheduled as a BackoffRunnable. Otherwise, it will be queued for later processing.

Specified by:

scheduleRetry in interface RunnableRetrier<T>

Parameters:

key - Runnable's key.

runnable - task to schedule.

Returns:

this.

hasQueuedRunnables

public boolean hasQueuedRunnables(T key)

Description copied from interface: RunnableRetrier

Return whether there are task pending to execute for an specific key.

Specified by:

hasQueuedRunnables in interface RunnableRetrier<T>

Parameters:

key - pending task's key.

Returns:

true/false.

isIdle

public boolean isIdle(T key)

Return whether the retrier is idle or there are tasks running or pending for execution for a specific key.

Parameters:

key - pending task's key.

Returns:

true if it is idle, false otherwise.

fastRecoveryUnstable

public FastRecoveryObserver fastRecoveryUnstable(BackoffRunnable runnable,
                                                 FastRecoveryConfiguration configuration)

Notifies the fast recovery BackoffRunnable is still Unstable. While BackoffRunnable remains Unstable, this object won't do anything. The BackoffScheduler will automatically reschedule the task.

Specified by:

fastRecoveryUnstable in interface FastRecoveryObserver

Parameters:

runnable - the Unstable BackoffRunnable.

configuration - the suggested delay and frequency if rescheduled along with its observers.

Returns:

this.

fastRecoveryStable

public FastRecoveryObserver fastRecoveryStable(BackoffRunnable backoffRunnable,
                                               FastRecoveryConfiguration configuration)

Notifies the fast recovery BackoffRunnable is now Stable. When BackoffRunnable returns to Stable state, the BackoffScheduler will remove the FastRecovery task. This object then, must schedule the next queued Runnable for that key, or dispose the scheduler if it was the last one.

Specified by:

fastRecoveryStable in interface FastRecoveryObserver

Parameters:

backoffRunnable - the Stable BackoffRunnable.

configuration - the suggested delay and frequency if rescheduled along with its observers.

Returns:

this.

fastRecoveryAbort

public FastRecoveryObserver fastRecoveryAbort(BackoffRunnable backoffRunnable)

Notifies the fast recovery BackoffRunnable is now with Error. When BackoffRunnable goes to Error state, the BackoffScheduler will remove the FastRecovery task and log the error. This object then, must schedule the next queued Runnable for that key, or dispose the scheduler if it was the last one.

Specified by:

fastRecoveryAbort in interface FastRecoveryObserver

Parameters:

backoffRunnable - the BackoffRunnable with Error.

Returns:

this.

dispose

public void dispose()

Specified by:

dispose in interface org.mule.runtime.api.lifecycle.Disposable

delayInitialScheduling

public static SchedulingConfiguration delayInitialScheduling(int initialDelay)

zeroDelayOnScheduling

public static SchedulingConfiguration zeroDelayOnScheduling()