Apex Async

Take control of your asynchronous Apex processes by using apex-async. The core idea is in composing asynchronous parts of the complex flow into the single chain and run all process from one place. This parts can be branched regarding the success or failure of the previous part.

How to

First, one or more classes which represent business logic should implement Queueable interface. Then create AsyncUnit instances from them. You can specify success flow or failure flow for every single unit. Concatenate async units with AsyncProcessor class and execute it. Please notes, if you've not specified failure flow, the processor will call the general on-error action. If general on-error action also hasn't specified exception won't be processed.

Make the chain

Join several (then) tasks in one process with AsyncProcessor and run it

AsyncProcessor aProcessor = new AsyncProcessor()
.then( 
  new AsyncUnit( new AccountUpdaterQueuable() ) 
)
.then( 
  new AsyncUnit( new ContactRemoteUpdaterQueuable() ) 
)
.then( 
  new AsyncUnit( new LoggerQueuable() ) 
)
.execute();

Create process with initial task

AsyncProcessor aProcessor = new AsyncProcessor(
  new AsyncUnit( new ContactRemoteUpdaterQueuable() )
)
.then( 
  new AsyncUnit( new AccountUpdaterQueuable() ) 
)
.execute();

Use fail function in AsyncProcessor for general handling errors

AsyncProcessor aProcessor = new AsyncProcessor()
.then( 
  new AsyncUnit( new AccountUpdaterQueuable() ) 
)
.fail( 
  new AsyncUnit( new LoggerQueuable() ) 
)
.execute();

Handle particular success and failure

Use fail function in AsyncUnit for particular handling errors

AsyncProcessor aProcessor = new AsyncProcessor()
.then( 
  new AsyncUnit( new AccountUpdaterQueuable() ) 
    .fail( new AccountUpdaterLogger() )
)
.fail( 
  new AsyncUnit( new LoggerQueuable() ) 
)
.execute();

Use success function in AsyncUnit for handling success from particular operation

AsyncProcessor aProcessor = new AsyncProcessor()
.then( 
  new AsyncUnit( new AccountUpdaterQueuable() ) 
    .success( new ContactRemoteUpdaterQueuable() )
    .fail( new AccountUpdaterLogger() )
)
.then( 
  new AsyncUnit( new OpportunityJobQueuable() ) 
    .success( new ContactRemoteUpdaterQueuable() )
    .fail( new OpportunityLogger() )
)
.execute();

Pass data from one job to another

For passing state from one job to another your classes should implement AsyncProcessor.Emitter and AsyncProcessor.Acceptor respectively. Example:

/* Create emitter class */
public class RemoteUpdater implements Queueable, AsyncProcessor.Emitter {
  private String responseBody;
  public void execute( QueueableContext context ) {
    /* Make callout */
    HTTPResponse res = http.send(req);
    this.responseBody = res.getBody();
  }
  public Object emit() {
    return this.responseBody;
  }
}

/* Create acceptor class */
public class APILogger implements Queueable, AsyncProcessor.Acceptor {
  private String responseBody;
  public void execute( QueueableContext context ) {
    /* Another business logic */
    insert new API_Log__c( 
      Response__c = this.responseBody
    );
  }
  public void accept( Object body ) {
    this.responseBody = (String) body;
  }
}

/* Create and execute flow */
AsyncProcessor aProcessor = new AsyncProcessor(
  new AsyncUnit( new RemoteUpdater() ) 
)
.then( 
  new AsyncUnit( new APILogger() )
)
.execute();