Think of it like DataDog 2.0: https://devnull-as-a-service.com/

I'm saying most logs are pointless to keep and would be better directed to /dev/null. Keep important transaction related logs and sample the rest.

The notion that every single log or metric across your entire technical architecture is worth keeping is one implanted by SaaS providers with a financial interest in naive engineers doing just that.

I agree. Many applications log for the hello of it and log unnecessary information.

We have concepts of debug, info, warn and error… but I think we need apps to be developed with the concept of log concern.

For example take sshd. For infosec they are interested in IP of failed attempts, operations might want to know connection failures, etc.

That sounds ridiculous, but on the other hand people trust their code to a service called bitbucket.