Step Functions: how to implement semaphores for state machines

Here at DAZN, we are migrat­ing from our lega­cy plat­form into a brave new world of microfron­tends and microser­vices. Along the way, we also dis­cov­ered the delights that AWS Step Func­tion has to offer, for exam­ple…

• flex­i­ble error han­dling and retry
• the under­stat­ed abil­i­ty to wait between tasks
• the abil­i­ty to mix auto­mat­ed steps with activ­i­ties that require human interven­tion

In some cas­es, we need to con­trol the num­ber of con­cur­rent state machine exe­cu­tions that can access a shared resource. This might be a busi­ness require­ment. Or it could be due to scal­a­bil­i­ty con­cerns for the shared resource. It might also be a result of the design of our state machine which makes it dif­fi­cult to par­al­lelise.

We came up with a few solu­tions that fall into two gen­er­al cat­e­gories:

1. Con­trol the num­ber of exe­cu­tions that you can start
2. Allow con­cur­rent exe­cu­tions to start, but block an exe­cu­tion from enter­ing the crit­i­cal path until it’s able to acquire a sem­a­phore (i.e. a sig­nal to proceed)

Control the number of concurrent executions

You can con­trol the MAX num­ber of con­cur­rent exe­cu­tions by intro­duc­ing a SQS queue. A Cloud­Watch sched­ule will trig­ger a Lamb­da func­tion to…

1. check how many con­cur­rent exe­cu­tions there are
2. if there are N exe­cu­tions, then we can start MAX-N exe­cu­tions
3. poll SQS for MAX-N mes­sages, and start a new exe­cu­tion for each

We’re not using the new SQS trig­ger for Lamb­da here because the pur­pose is to slow down the cre­ation of new exe­cu­tions. Where­as the SQS trig­ger would push tasks to our Lamb­da func­tion eager­ly.

Also, you should use a FIFO queue so that tasks are processed in the same order they’re added to the queue.

Block execution using semaphores

You can use the Lis­tEx­e­cu­tions API to find out how many exe­cu­tions are in the RUNNING state. You can then sort them by start­Date and only allow the eldest exe­cu­tions to tran­si­tion to states that access the shared resource.

Take the fol­low­ing state machine for instance.

The Only­One­Shall­RunA­tOne­Time state invokes the one-shall-pass Lambda func­tion and returns a proceed flag. The Shall Pass? state then branch­es the flow of this exe­cu­tion based on the proceed flag.

OnlyOneShallRunAtOneTime:  Type: Task  Re