akashchouhan16 / potd-node.js-service Goto Github PK

• Currently, the server makes a query to the DB with the iterator value to fetch the problem of the day, for every request made by the client on /problemoftheday.
• This is not at all scalable and it also leads to very high latencies to process each request. Caching the problem of the day to the server memory itself with the first request. ~600-1200ms. Further get requests made by the client could directly access the problem of the day from the server cache, hence reducing latency to less than ~50-60ms.

• Add Caching onto the server, to avoid network requests to DB for every client-side request.

• Current version is running on Heroku's free tier instance
• Migrate to DigitalOcean droplet or any low-latency deployment service.

• Implement API rate limiter for /problemoftheday api
• Expected Config

const limitConfig = {
    windowMs: 5*60*1000,                    // 5min window
    max: 7                                 //max no of requests per window
};

• The server should utilize multiple CPUs if present on the system, to run concurrent requests.

Note: Could be achieved with cluster & os module in Node.

• The Problem Of The Day should be picked randomly from the DB, instead of sequentially.
• Updates should be scheduled on a recurring basis using Cron.