[nathcd]

> To folks saying to “Just use server logs”, it’s simply too hard to connect the dots unless you really know what you’re doing.

No it's not, just stop overarchitecting your stuff.

> Digging through my nginx logs to find that the majority of my users are dropping off at the second step of sending an invoice would be nearly impossible. Oh I could connect those logs to my api’s database, now I need a place to aggregate that data together. Maybe I could send all of that data together to elasticsearch and query with kibana. Now I need to learn how To set those things up.

You can log requests/responses however you want on your server, it doesn't have to be an nginx log. Just put whatever request data that matters to you into your database immediately after you send a response.

> Or I could use Matomo for some analytics real quick

"Not being a jerk to my website visitors is too hard."

[CathedralBorrow]

I don't know, it seems like an arbitrary line to me. If I logged everything about your visit in a file, then wrote some code to analyze and reconstruct a picture of you as a user, would it be OK with you because I call it a server log?

[nathcd]

When I hit <enter> in a URL bar, or click on a link to a webpage, or fill out a form on the web, I know that I'm sending a request to a server. That's why I'm taking that action in the first place. But if you send me a program in a webpage to collect data about me, I'm not trying to send you that data, you're doing it behind my back.

You can still certainly be a jerk with data you collect in your server logs. (For example, I'd be pretty angry if you sold your log data to google or some other advertiser.) But your server is your computer, and the data in server logs is the data making up the HTTP requests that were explicitly sent to your server. That's your data, not mine.