Can you share a bit more about the architecture and stack? And a few more details such as Roughly how many TPS do you serve from this setup? Does it scale well? Are response times consistent?

It's all written in Go + Postgres. Nothing else. Not sure what TPS is.

No problem with scaling.

Stack Overflow runs on a single machine. Use a fast compiled language and a simple stack. No need in $6k/mo AWS bills which are the fashion these days.

Your response gives me zero confidence that you know what you're talking about. Having "hundreds of thousands" of users tells us nothing (which is why TPS was brought up). Also if you do have a high concurrent user count then there's no way there have been "no issues scaling". There are always scale points in systems that actually grow.

Since 2015, no issues. You can choose not to believe me.

You're choosing not to try to convince anyone that what you're saying is true. Why bother posting at all, in that case?

Just show a metric/screenshot, it isn't hard to prove what you are saying.

I absolutely choose not to believe people that show zero proof XD

Stack Overflow is a read heavy website. They can get away with caching most of their requests in POPs around the world which will deal with the vast majority of traffic.

Having thousands of users is a meaningless metric for web operations. Are those concurrent users? Are those read/write heavy users? Are those mixed usage users? There are a lot of qualifiers that need to be attached to that question to allow any meaningful discourse.

SO doesn't run on a single machine, they have a very detailed blog post around their architecture.

Ok, 2 web servers, 2 db servers back in 2013, one of the most visited sites out there.

https://nickcraver.com/blog/2013/11/22/what-it-takes-to-run-...

1 - you said it as was it was current, it is not.

2 - still, in 2013 is not 1 server. It's also more than you said now ( add a tag engine server, elastic search, redis, load balancer) and an insane amount of memory on a lot of those servers. Even then, the stack exchange network is more than that.

Saying 1 server, as you did, is not correct and it never was :)

Transactions Per Second.

Always looking to optimize on cost for AWS. But it's a tradeoff of something or the other. Light Sail sounds like a great solution for some use-cases though.

We run this for a Global/North American airline and easily hit 1k TPS just on API and event processing traffic. The spikes can show up anywhere in the multitude of processes due to various business events (high reservation traffic to irregular operations due to weather, pandemics etc.) Performance and reliability wise, serverless has been a very predictable consistent experience. Costs are another story.

I'm about to try the same thing, but with crystal + postgres + nginx + cloudflare + serverless, so we'll see.

Oh wow, does lightsail come with postgres? Is it like a PaaS on top of AWS? Also TPS means transactions per second

LightSail is like DigitalOcean/Linode/Vultr, simple fixed cost VMs and DBs only.

Does the $3.5 per month include DB as well?

You get a system that you can run anything on.

You can run a database in the same instance, but all your tables have to fit in 0.6 TB or less, and hardware failures will cause downtime and possibly corruption.

DB & API on the same instance doesn't sound ideal

"TPS" is usually transactions per second.