The app is a simple microservice for rating conference talks. It provides a web API to track the ratings of the talks in real time, storing the results in a PostgreSQL database and a Redis cache, and using Redpanda as a broker for the event stream. Finally, it will use an AWS lambda function to calculate some statistics about the ratings of a talk.
When a rating is submitted, we must verify that the talk for the given ID is present in the database.
The database of choice is PostgreSQL, accessed with jackc/pgx PostgreSQL Driver.
Check internal/talks/repo.go.
The application stores the ratings in a Redis store with redis/go-redis Redis client.
Check internal/ratings/repo.go.
The application uses ES/CQRS to materialize the events into the state. Redpanda acts as a broker and the twmb/franz-go Kafa client.
Check internal/streams/broker.go.
The application uses AWS lambda functions to calculate some rating statistics of a talk. The lambda functions are invoked by the application when the ratings for a talk are requested, using HTTP calls to the function URL of the lambda.
Check internal/ratings/lambda_client.go.
The Go lambda application will live in the lambda-go directory. It is a simple lambda function that receives the talk ID and returns some statistics of the talk.
We'll start adding the code for the lambda in the Step 7.
The API is a Gin REST controller and exposes three endpoints:
POST /ratings { "talkId": ?, "value": 1-5 }to add a rating for a talkGET /ratings?talkId=?to get the histogram of ratings of the given talkGET /returns metadata about the application, including the connection strings to all the backend services (PostgreSQL, Redis, Redpanda).
Check internal/app/handlers.go.