Is real google bot like "fetch"or more like "fetch & render"?

[max.favilli]

In GWT we have two options to mimic googlebot visits, "fetch" and "fetch and render", but when the real googlebot visit a page, is he behaving like the former or the latter?

I can see fetch does fetch only the html, while fetch and render does fetch .js and .css as well.

But what does the real googlebot does?

I have checked the web server logs, and I can see the real googlebot sometimes request the .js files too, but not every time it visit a page, sometimes it does, sometimes it does not.

Has anyone figured out when googlebot actually request javascript files?

[max.favilli]

That was my guess, that google somehow cache the .js files, so I changed one javascript file name to a random generated name. Did it last night, I will recheck logs in a couple of days to see what happen.

[RyanPurkey]

Googlebot is behaving like the latter, "Googlebot constantly crawls and renders URLs across the entire web..." from https://support.google.com/webmasters/answer/6066467, but has the option of streamlining the process by not rendering every time. For example, if many sites are referencing the same centrally hosted code base of js, googlebot doesn't need to constantly recheck that js every time it comes across the link.