Momento MongoDB Read-Through Cache Demo

This TypeScript application runs in Lambda and uses Momento's serverless cache as a read-through cache for MongoDB. To deploy the application yourself you'll need a running MongoDB serverless/cluster instance, a Momento account, and an AWS account to deploy the app. Then copy .env.template to .env and fill in each line. You will need:

• AWS_ACCOUNT_ID= the numeric AWS account ID to deploy to
• COLLECTION_NAME= the name of both your Momento cache and the MongoDB collection where your players will be stored. Create the Momento cache with momento cache create myGame and a MongoDB collection with the same name.
• MONGODB_URI= a URI from MongoDB Atlas like mongodb+srv://USER:PASSWORD@cluster....mongodb.net
• MOMENTO_AUTH_TOKEN= the JWT token to access Momento and store the lookaside cache

With the .env file ready, ship the app with npm run deploy and the CDK will deploy and wait for the AWS resources needed to run the app.

Caching Principles

Applications often have globally visible leaderboards, reports, or rankings that don't change frequently but are relatively expensive to query. Caching those commonly queried but long-lived values makes more system resources available for other requests. In this app, we cache the values for find, count, and distinct queries for 60 seconds. Caching is the ultimate "It Depends" in application development. Sometimes real-time data is important, other times a leaderboard might remain static for hours or days.

To make it easy for developers to use cached values, we add a shim to MongoDB (via the Mongoose SDK) globally so any time a query is made we replicate the result to Momento.

The API

This application implements a simple game - one you can win by knowing how to use curl. To add one point to your score, make a POST request like: curl -XPOST 'https://YOUR_API_URL.amazonaws.com/players?game=tennis&name=Alice'

Your score will be sent back, along with the game you're participating in. There's no limit on the number of games you can play, or how many times you can increment your score. To see how you're doing, grab the leaderboard with a GET request to the same curl -s https://YOUR_API_URL.amazonaws.com/players\?game\=tennis | jq . path (with JQ to make the output prettier).

{
  "elapsed": 28,
  "game": "tennis",
  "result": [
    {
      "_id": "641488e28ac88ffda14464fe",
      "game": "tennis",
      "name": "Alice",
      "score": 4
    },
    {
      "_id": "641488d28ac88ffda144550d",
      "game": "tennis",
      "name": "Candace",
      "score": 3
    },
    {
      "_id": "641489078ac88ffda1448ae7",
      "game": "tennis",
      "name": "David",
      "score": 3
    },
    {
      "_id": "641489038ac88ffda14486fd",
      "game": "tennis",
      "name": "Bob",
      "score": 1
    }
  ]
}