Developing CloudStatus, an Alexa Skill to Query AWS Service Status -- an interview with Kira Hammond by Eric Hammond

Interview conducted in writing July-August 2016.

[Eric] Good morning, Kira. It is a pleasure to interview you today and to help you introduce your recently launched Alexa skill, “CloudStatus”. Can you provide a brief overview about what the skill does?

[Kira] Good morning, Papa! Thank you for inviting me.

CloudStatus allows users to check the service availability of any AWS region. On opening the skill, Alexa says which (if any) regions are experiencing service issues or were recently having problems. Then the user can inquire about the services in specific regions.

This skill was made at my dad’s request. He wanted to quickly see how AWS services were operating, without needing to open his laptop. As well as summarizing service issues for him, my dad thought CloudStatus would be a good opportunity for me to learn about retrieving and parsing web pages in Python.

All the data can be found in more detail at But with CloudStatus, developers can hear AWS statuses with their Amazon Echo. Instead of scrolling through dozens of green checkmarks to find errors, users of CloudStatus listen to which services are having problems, as well as how many services are operating satisfactorily.

CloudStatus is intended for anyone who uses Amazon Web Services and wants to know about current (and recent) AWS problems. Eventually it might be expanded to talk about other clouds as well.

[Eric] Assuming I have an Amazon Echo, how do I install and use the CloudStatus Alexa skill?

[Kira] Just say “Alexa, enable CloudStatus skill”! Ask Alexa to “open CloudStatus” and she will give you a summary of regions with problems. An example of what she might say on the worst of days is:

“3 out of 11 AWS regions are experiencing service issues: Mumbai (ap-south-1), Tokyo (ap-northeast-1), Ireland (eu-west-1). 1 out of 11 AWS regions was having problems, but the issues have been resolved: Northern Virginia (us-east-1). The remaining 7 regions are operating normally. All 7 global services are operating normally. Which Amazon Web Services region would you like to check?”

Or on most days:

“All 62 regional services in the 12 AWS regions are operating normally. All 7 global services are operating normally. Which Amazon Web Services region would you like to check?”

Request any AWS region you are interested in, and Alexa will present you with current and recent service issues in that region.

Here’s the full recording of an example session:

[Eric] What technologies did you use to create the CloudStatus Alexa skill?

[Kira] I wrote CloudStatus using AWS Lambda, a service that manages servers and scaling for you. Developers need only pay for their servers when the code is called. AWS Lambda also displays metrics from Amazon CloudWatch.

Amazon CloudWatch gives statistics from the last couple weeks, such as the number of invocations, how long they took, and whether there were any errors. CloudWatch Logs is also a very useful service. It allows me to see all the errors and print() output from my code. Without it, I wouldn’t be able to debug my skill!

I used Amazon EC2 to build the Python modules necessary for my program. The modules (Requests and LXML) download and parse the AWS status page, so I can get the data I need. The Python packages and my code files are zipped and uploaded to AWS Lambda.

Fun fact: My Lambda function is based in us-east-1. If AWS Lambda stops working in that region, you can’t use CloudStatus to check if Northern Virginia AWS Lambda is working! For that matter, CloudStatus will be completely dysfunctional.

[Eric] Why do you enjoy programming?

[Kira] Programming is so much fun and so rewarding! I enjoy making tools so I can be lazy.

Let’s rephrase that: Sometimes I’m repeatedly doing a non-programming activity—say, making a long list of equations for math practice. I think of two “random” numbers between one and a hundred (a human can’t actually come up with a random set of numbers) and pick an operation: addition, subtraction, multiplication, or division. After doing this several times, the activity begins to tire me. My brain starts to shut off and wants to do something more interesting. Then I realize that I’m doing the same thing over and over again. Hey! Why not make a program?

Computers can do so much in so little time. Unlike humans, they are capable of picking completely random items from a list. And they aren’t going to make mistakes. You can tell a computer to do the same thing hundreds of times, and it won’t be bored.

Finish the program, type in a command, and voila! Look at that page full of math problems. Plus, I can get a new one whenever I want, in just a couple seconds. Laziness in this case drives a person to put time and effort into ever-changing problem-solving, all so they don’t have to put time and effort into a dull, repetitive task. See

But programming isn’t just for tools! I also enjoy making simple games and am learning about websites.

One downside to having computers do things for you: You can’t blame a computer for not doing what you told it to. It did do what you told it to; you just didn’t tell it to do what you thought you did.

Coding can be challenging (even frustrating) and it can be tempting to give up on a debug issue. But, oh, the thrill that comes after solving a difficult coding problem!

The problem-solving can be exciting even when a program is nowhere near finished. My second Alexa program wasn’t coming along that well when—finally!—I got her to say “One plus one is eleven.” and later “Three plus four is twelve.” Though it doesn’t seem that impressive, it showed me that I was getting somewhere and the next problem seemed reasonable.

[Eric] How did you get started programming with the Alexa Skills Kit (ASK)?

[Kira] My very first Alexa skill was based on an AWS Lambda blueprint called Color Expert (alexa-skills-kit-color-expert-python). A blueprint is a sample program that AWS programmers can copy and modify. In the sample skill, the user tells Alexa their favorite color and Alexa stores the color name. Then the user can ask Alexa what their favorite color is. I didn’t make many changes: maybe Alexa’s responses here and there, and I added the color “rainbow sparkles.”

I also made a skill called Calculator in which the user gets answers to simple equations.

Last year, I took a music history class. To help me study for the test, I created a trivia game from Reindeer Games, an Alexa Skills Kit template (see That was a lot of fun and helped me to grow in my knowledge of how Alexa works behind the scenes.

[Eric] How does Alexa development differ from other programming you have done?

[Kira] At first Alexa was pretty overwhelming. It was so different from anything I’d ever done before, and there were lines and lines of unfamiliar code written by professional Amazon people.

I found the ASK blueprints and templates extremely helpful. Instead of just being a functional program, the code is commented so developers know why it’s there and are encouraged to play around with it.

Still, the pages of code can be scary. One thing new Alexa developers can try: Before modifying your blueprint, set up the skill and ask Alexa to run it. Everything she says from that point on is somewhere in your program! Find her response in the program and tweak it. The variable name is something like “speech_output” or “speechOutput.”

It’s a really cool experience making voice apps. You can make Alexa say ridiculous things in a serious voice! Because CloudStatus started with the Color Expert blueprint, my first successful edit ended with our Echo saying, “I now know your favorite color is Northern Virginia. You can ask me your favorite color by saying, ‘What’s my favorite color?’.”

Voice applications involve factors you never need to deal with in a text app. When the user is interacting through text, they can take as long as they want to read and respond. Speech must be concise so the listener understands the first time. Another challenge is that Alexa doesn’t necessarily know how to pronounce technical terms and foreign names, but the software is always improving.

One plus side to voice apps is not having to build your own language model. With text-based programs, I spend a considerable amount of time listing all the ways a person can answer “yes,” or request help. Luckily, with Alexa I don’t have to worry too much about how the user will phrase their sentences. Amazon already has an algorithm, and it’s constantly getting smarter! Hint: If you’re making your own skill, use some built-in Amazon intents, like AMAZON.YesIntent or AMAZON.HelpIntent.

[Eric] What challenges did you encounter as you built the CloudStatus Alexa skill?

[Kira] At first, I edited the code directly in the Lambda console. Pretty soon though, I needed to import modules that weren’t built in to Python. Now I keep my code and modules in the same directory on a personal computer. That directory gets zipped and uploaded to Lambda, so the modules are right there sitting next to the code.

One challenge of mine has been wanting to fix and improve everything at once. Naturally, there is an error practically every time I upload my code for testing. Isn’t that what testing is for? But when I modify everything instead of improving bit by bit, the bugs are more difficult to sort out. I’m slowly learning from my dad to make small changes and update often. “Ship it!” he cries regularly.

During development, I grew tired of constantly opening my code, modifying it, zipping it and the modules, uploading it to Lambda, and waiting for the Lambda function to save. Eventually I wrote a separate Bash program that lets me type “edit-cloudstatus” into my shell. The program runs unit tests and opens my code files in the Atom editor. After that, it calls the command “fileschanged” to automatically test and zip all the code every time I edit something or add a Python module. That was exciting!

I’ve found that the Alexa speech-to-text conversions aren’t always what I think they will be. For example, if I tell CloudStatus I want to know about “Northern Virginia,” it sends my code “northern Virginia” (lowercase then capitalized), whereas saying “Northern California” turns into “northern california” (all lowercase). To at least fix the capitalization inconsistencies, my dad suggested lowercasing the input and mapping it to the standardized AWS region code as soon as possible.

[Eric] What Alexa skills do you plan on creating in the future?

[Kira] I will probably continue to work on CloudStatus for a while. There’s always something to improve, a feature to add, or something to learn about—right now it’s Speech Synthesis Markup Language (SSML). I don’t think it’s possible to finish a program for good!

My brother and I also want to learn about controlling our lights and thermostat with Alexa. Every time my family leaves the house, we say basically the same thing: “Alexa, turn off all the lights. Alexa, turn the kitchen light to twenty percent. Alexa, tell the thermostat we’re leaving.” I know it’s only three sentences, but wouldn’t it be easier to just say: “Alexa, start Leaving Home” or something like that? If I learned to control the lights, I could also make them flash and turn different colors, which would be super fun. :)

In August a new ASK template was released for decision tree skills. I want to make some sort of dichotomous key with that.

[Eric] Do you have any advice for others who want to publish an Alexa skill?


  • Before submitting your skill for certification, make sure you read through the submission checklist.

  • Remember to check your skill’s home cards often. They are displayed in the Alexa App. Sometimes the text that Alexa pronounces should be different from the reader-friendly card content. For example, in CloudStatus, “N. Virginia (us-east-1)” might be easy to read, but Alexa is likely to pronounce it “En Virginia, Us [as in ‘we’] East 1.” I have to tell Alexa to say “northern virginia, u.s. east 1,” while leaving the card readable for humans.

  • Since readers can process text at their own pace, the home card may display more details than Alexa speaks, if necessary.

  • If you don’t want a card to accompany a specific response, remove the ‘card’ item from your response dict. Look for the function build_speechlet_response() or buildSpeechletResponse().

  • Never point your live/public skill at the $LATEST version of your code. The $LATEST version is for you to edit and test your code, and it’s where you catch errors.

  • If the skill raises errors frequently, don’t be intimidated! It’s part of the process of coding. To find out exactly what the problem is, read the “log streams” for your Lambda function. To print debug information to the logs, print() the information you want (Python) or use a console.log() statement (JavaScript/Node.js).

  • It helps me to keep a list of phrases to try, including words that the skill won’t understand. Make sure Alexa doesn’t raise an error and exit the skill, no matter what nonsense the user says.

  • Many great tips for designing voice interactions are on the ASK blog.

  • Have fun!

In The News

Amazon had early access to this interview and to Kira and wrote an article about her in the Alexa Blog:

14-Year-Old Girl Creates CloudStatus Alexa Skill That Benefits AWS Developers

which was then picked up by VentureBeat:

A 14-year-old built an Alexa skill for checking the status of AWS