RR 323: Queuing and Amazon SQS with Kinsey Ann Durham

This episode of Ruby Rogues features panelists Charles Max Wood, Dave Kimura, and Eric Berry. Special guest Kinsey Ann Durham joins to talk about queuing and Amazon SQS. Tune in to learn more!

[00:01:19] Kinsey Ann Durham

Kinsey writes code for a company called Go Spot Check. She is always a lead mentor in a San Francisco based company called Bloc.

[00:02:50] Background on Amazon SQS

Go Spot Check is using Amazon SQS on a smaller scale. Kinsey thinks it is sasy to use. She recommends using something like Amazon SQS or even RabbitMQ. It has provided the company with the ability to explore different architecture patterns and tools.

[00:04:50] Can you talk a little about your company and what led to using Amazon SQS?

Go Spot Check is a start up in Denver. They focus on recording and data collection for big companies that need to know what is happening in retail, grocery stores, and bars. The focus is on alcohol and retail brands. The company analyzes the data collected that previously held no insight. Go Spot Check is currently moving into a computer vision aspect. Kinsey works off a separate service off of main aspect of Go Spot Check.

[00:06:46] What does your stack look like? Is it built off Ruby?

Yes, it is a Rails API only. The computer vision is done in Python.

[00:08:45] Are you feeding the images through the queue? How does the queuing fit in?

Started using Amazon SQS because they wanted to have a more decoupled way of developing. This allowed them to decide the contract between the two services and decide what they wanted it to look like up front. Kinsey describes that it is easy to create fake messages for testing with Amazon SQS. Image data is sent back and forth through the queue. The company does a lot of planograms. Information is taken from that data and posted onto a queue from the machine learning side of things. On the Rail side of things, the data can be picked up in API and sent back to the main app.

[00:10:50] Does it accept binary data in the queue?

It does not send actual images. All comparison data that has been processed is sent from the machine learning aspect side of things. An article has been published that shows that people do send images in the queue.

[00:11:35] Do you use SQS in parallel with SNS (Simple Notification Service)?

Kinsey says that they haven’t used SNS. This is because there hasn’t been a need. They are using it to post messages to communicate between different services.

[00:12:40] What point would you need to consider a SQS over a Sidekick?

Kinsey didn’t look into using Sidekick; she was excited to use SQS. She wanted to try it out and see if it was easy to use. Thought it would be more complex than it has been. She enjoys the free features of Amazon such as message visibility and timeout, which is handled by them. It can be customized and two different queues can be used.

[00:16:15] How do you write the workers for an SQS queue?

Kinsey has a plain Ruby object in the API that she can reuse with any queue. There are three queues in the company.

[00:19:45] Are there any other uses for queues and SQS?

Kinsey hasn’t come across any personally but she is sure there are some.

[00:23:40] What if you’re someone who is new? Where would you recommend they get started?

Suggest getting started with SQS Amazon, SQS documentation. Can get up to speed quickly. Amazon SQS is easy to get up and running. Kinsey is tailoring her Ruby Dev Summit talk to people who are new.

[00:30:35] How do you go about mentoring?

Kinsey loves mentoring. Developers have side projects or freelance work, but Kinsey likes to mentor because she feels like she makes a difference while continuing to learn. An important part of mentorship is giving support. This support level to students’ means not only offering students help with technical skills. Her goal is to build a well-rounded developer: someone who will be a great team member and people will