There are plenty of logging stacks to choose from. I've used Elastic a lot and it has improved a lot. You can spin up a cheap cluster for under 200$ and start instrumenting your servers to send stuff to it. You'll want to set up index life cycle management to ensure you don't run out of disk space on your cluster. You scale by throwing more money at it. Basically you need to think in terms of millions of messages per day and retention periods. A 200$ cluster should be able to retain tens of millions of messages.

That's how you get started. There are plenty of tools on this stack to do APM, security auditing, request logging, etc. If you are using a decent application server stack that produces metrics, it can handle those too.