Hacker News Reader

Hey thank you! Yes totally, what would you like to know?

The backend is python/starlette/marshmallow running serverless on lambda behind apigateway. It uses many other lambda/sqs functions, dynamodb, iot for websockets.

The frontend is an ember.js octane app with ssr via fastboot.js in cloudfront lambda@edge.

View on HN

What is the approximate load and cost?

It's $150/month right now for 3 aws accounts (parent/staging/prod).

If you increased your usage metrics by 10, 100, 1000 what does that do to costs? How many users or messages an hour would make it untenable for you to keep hosting?

To be honest I'm not sure off the top of my head, lambda/dynamo/s3/iot scale quite a bit. Opensearch is one of the most costly services but I have it behind apigateway cache.

I'm sure those services can scale but what I wonder about is if the cost of scaling will become prohibitive. If your posts per day increase by three orders of magnitude will your costs?

That seems insanely expensive for something that looks like it gets maybe a post every few minutes.

Have you tried running it on, like, a $20/month VPS?

Reply on news.ycombinator.com