I wanted to ask if this is intended or an unintended consequence of something.

I used GPT to ask about plugin options, if something existed to do xyz kind of thing. GPT told me about the welcome plugin and gave me the URL to go read about it. (https://doc.dovecot.org/2.4.0/core/plugins/welcome.html?utm_source=chatgpt.c...)

However, the URL with the ?utm... was being denied with a "400 Bad Request - nginx" error. If i removed the ?utm... part from the URL then the page would load fine in the browser. Put the ?utm... back and i'd get the error again.

Why is the documentation blocking URL's with gpt referral links? It's not stopping scraping, only us humans are getting blocked by it.