Is nobody using log tables? Pretty much every time I touch something in my db, there's a log call that records who did it, when, IP, URL and a (JSON) snapshot o

Is nobody using log tables? Pretty much every time I touch something in my db, there's a log call that records who did it, when, IP, URL and a (JSON) snapshot of the changed record, which in a pinch can be used for undelete. It's surprisingly manageable. I mean, yes, it's definitely the largest table in the db, but: 1. it's well worth it 2. most of the stuff in it isn't the main scenario above (a human does something and I record the change) but various automated processes I also want to track, like API calls. which leads to: 3. it's easy to prune - both in time period kept, and by selectively deleting the automated stuff earlier But it mostly helps by localizing things. It's just one meta-data log table, and everything related to logging actions is there. Not very elegant to keep adding fluff fields to every table, like "add_date" or "deleted_at". When I decided I want to also track the URL of the request I had to change things in just one place, and now I have it for every action everywhere. Note: don't fall into the "everything is a nail" mistake. Some other dedicated log tables may be necessary, for high-volume or distinct stuff. I also have a mail_log, a sms_log and a separate table for events coming from mobile users (like location history).