Cloud resume Challenge (AWS version)

The project was built based on instructions from the website for the Cloud resume challenge.

I will fashion the effort I put into this project into steps for brevity:

• Create an s3 bucket.
• Create a cloudfront distribution to serve static files from aforementioned s3 bucket.
• Write configuration for the dynamoDB table to store a value (number of visits to website).
  • A "Composite key" is what dynamoDB needs to index the data. Technically, one could do without this, but this was the easiest option, as there would be other complications in using a single key which I did not explore.
  • The composite key is formed from the HASH and RANGE keys.
  • In my case, the data type indexed by either key was an integer, but one can change that.
• Create python script to "interface" with the table (using the boto3 library).
  • The response from the script is very important for the forthcoming API, as I will mention in a subsequent point.
• Configure lambda function to utilise the script.
  • Every lambda instance passes the variables event and context to the function running inside it. I had to make sure that my function, at the least, accepted the two variables (even though I didn't get around to using them).
  • For the lambda to actually run this code, it needs to be able to access it, and basically read it.
    • If one has some simple code, zip is a good option.
    • But that wouldn't work in my case, so I had to configure an s3 bucket, along with the requisite permissions for the lambda function to access the file inside the bucket and read from that.
• Write configuration for API Gateway to utilise mentioned lambda function and get the value from the database table.
  • Without the correct response with the right parameters (this was an HTTP API so required some specific headers in the response, probably different for REST APIs), the output would be NULL/None.
• The next problem I faced was: to call the API, I would need to know the URL.
  • The problem being that AWS assigns an unique ID/combination of characters to the API when it is created.
  • Thus, I needed a way to scrape the API URL and then invoke it.
• Thus, create another lambda function with more code, this time to list APIs in an account (the credentials were served using environment variables, which are destroyed when the lambda instance is decommissioned).
  • This lambda function would then write the API URL to a file.
  • Then, this API would be called using the API from the file.
  • Of course, this necessitated the need for another file and more permissions.
• The Shell scripts utilised should work on POSIX systems (inasmuch assuming that Debian's Dash can be trusted to maintain compatibility).
  • The first shell script s3run.sh invokes the cloudformation template for the storage bucket (holds the code for both lambda functions).
  • The second script, run.sh invokes the main template, template.yaml, creates the entire infrastructure (other than that of the storage-bucket, which was done by s3run.sh).
    • It then uploads the static website files, creates lambda environment variables.
    • The various flags that can be used with this script enable it to both deploy and delete the main template, both of which call the correct, related flag of the template s3run.sh from itself.
  • Both scripts have a "help" function to list their capabilities.

---

I have used many resources from around the web for this project, this is a great resource if anyone wants a reference for the project (note that the author uses SAM whilst I have exclusively stuck with cloudformation):

• Video series refererred to here: https://www.youtube.com/playlist?list=PLEk97Q5Nj5oesA1WNk7DzaUpZUnCsQFVQ
  • Github link for ^: https://github.com/openupthecloud/cloud-resume-challenge