Need suggestion about Server configuration?

As @seanmcgary says, this is so dependent on the webserver, the architecture of the application and the type of work your application is doing.

Most good hosting providers will easily let you scale up your instances RAM, CPU and storage requirements with just a few clicks. I would suggest starting small and seeing how the webserver handles the load of 2-3k requests per minute, and then scale up as need be.

Whether to use redis or memcache is again really dependent on the situation. They are just tools to get a job done. What is the problem that needs to be solved? - Then I might be able to help on this one.

As for serverless - I personally find it slows down development and makes things more complex - but I haven't played around with serverless architecture much so it is likely inexperience.

If you really need a quick and dirty answer - Im quite partial to DigitalOcean or Linode. I've never used memcache but love redis and it's widely supported by hosts and software languages.

This is wildly going to depend on a number of factors:

• What language/framework/webserver are you using?
• What is happening during each request? A route that calls directly to redis and returns a result is much faster/lighter weight than one that accepts a file upload
• Are you actually serving 2k requests per minute currently and need to scale, or are you guessing at how much traffic you're going to receive?

Id say start on something small and see how it goes. 512MB of memory and 1-2 vCPUs should get you pretty far if you're just getting started.

Redis or Memcache?

Id say Redis. Not to say that Memcache is bad, but over the years Redis has definitely won the popularity contest and offers many more features.

Which hosting provider should I choose for your suggested configuration?

Im partial to AWS since Ive used it professionally and for my own side projects. They offer both compute hosting and hosted versions of Redis and Memcache if you dont feel like managing them yourself.