Daily Advice For The Engineer (DAFTE)

DAFTE (or Dafte) is a learning project by Ben to explore building a Google Cloud Function. Or at least that's how it started out.

After building out the basic use case as a GC Function, it seemed worthwhile to consider that a single use approach such as a GC Function might not be the most efficient way to handle requests at various scales/capacities. To be fair to both approach, the project was adapted to either run as a deployed GC Function, or as a Standalone Spark microservice.

Running Locally

As a Google Cloud Function

To run DAFTE locally in the same way that it will be invoked as a GC Function:

./gradlew runFunction -Prun.functionTarget=dafte.Dafte

Then make requests to localhost:8080/dafte

As a Spark Microservice

To run DAFTE locally using the Spark microservice framework:

./gradlew run

Then make requests to localhost:4567

In Docker via Spark (Experimental, at Best)

There is a Dockerfile that very loosely defines a way that Spark can technically be started and run, in a rather not-so-production-ready way. To build and run:

docker build -t dafte .
docker run -p 4567:4567 dafte

Then make requests to localhost:4567

Supported Query Parameters

DAFTE Supports the following query parameters to alter the default randomized advice giving:

• name=<name>: Specify a name for the individual being advised. This creates a date and name based seed that will return the same advice for the span of the calendar day.
  • Example: localhost:4567?name=yukon-gold
• shape=raw: Specify the response shape
  • Example: localhost:4567?shape=raw
  • Supported Shapes:
    • raw: Returns a plain text string with only the generated advice
    • json: Return json in the following format:

      • {
          "preparedFor": "yukon-gold",  // Only when `name` is specified
          "date":"2021-06-18",
          "advice":"The advice is in the journey."
        }

Deploying as a Google Cloud Function

A more comprehensive guide can be found within Google's documentation.

If you have set up the following (instructions available at the above link):

• gcloud sdk installed locally and in your path
• a GCP Project created
• Billing enabled for your Project
• The Cloud Functions and Cloud Builds API enabled

Then you can deploy the DAFTE function either using the gcloud function command:

gcloud functions deploy dafte --entry-point dafte.Dafte --runtime java11 --trigger-http --memory 512MB --allow-unauthenticated

or by using the shorthand deploy-gcfunction.sh script, which will run that exact command, you just won't have to remember all of it.