Google Faviconbot

[lucy24]

Not an ID* but a question: has anyone got a comprehensive list of what this UA does, or rather who it works for? I know of two:

-- in GSC, the faviconbot is responsible for the icons on your dropdown list of sites
-- in Google+ or Profile or whatever it's called these days, the favicon provides the icons if you name websites you're associated with.

But I've found visits that can't be explained by either of these. They're not experimenting with including the favicon on SERPs are they?


* Last time I looked, it's:

Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/49.0.2623.75 Safari/537.36 Google Favicon

[keyplyr]

You're not giving the range. Check to see if it's coming from a Google crawl range. Has to be a valid crawl range, not just a range owned by Google. Otherwise it could be anything, from anyone.

But assuming this UA is Google, besides the tasks mentioned, its purpose could very well have been expanded to include new features of the upcoming Mobile-First Index [webmasterworld.com]

[lucy24]

The faviconbot has never come from the crawl range as such (66.249.64-79); it uses the associated .80-95 "miscellaneous Googloid functions" range alongside 66.102.whatever-it-is. There's no difference in the two--I just re-checked--so I have two consider both equally legitimate.

The headscratchers are the ones where the favicon request is preceded not by the root (as in the two functions I know about) but by some other interior file, rarely even an image. Either way, it's one file + favicon.

new features of the upcoming Mobile-First Index

I've been reading that thread, but I don't remember anything about icons.

[keyplyr]

Still waiting for the ip address. I can do much without that.

I've been reading that thread, but I don't remember anything about icons.

and I said...

its purpose could very well have been expanded to include new features of the upcoming Mobile-First Index

Emphasis on "could" since the Mobile-First Index is not implemented yet, no one knows.

- - -

There is a bookmarker named Google Favicon that's an Add-On for various browsers:
[chrome.google.com...]
[addons.mozilla.org...]

[keyplyr]

OK... just got back from the hunt (whew!)

66.102.6.86 - - [07/Jul/2017:00:54:59 -0700] "GET /example.html HTTP/1.1" 403 9814 "-" "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/49.0.2623.75 Safari/537.36 Google Favicon"

This is Google Proxy:
66.102.0.0 - 66.102.15.255
66.102.0.0/20

This is not a Google crawl range. These ranges are owned by Google, but are leased to anyone for any purpose (think AWS.) I see a lot of app upstarts using these ranges.

Personally, since these are not actual humans using these proxy ranges, I block them, allowing exceptions.

[lucy24]

66.249.80.7 - - [07/Jul/2017:18:00:05 -0700] "GET / HTTP/1.1" 301 562 "-" "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/49.0.2623.75 Safari/537.36 Google Favicon" 
66.102.7.142 - - [07/Jul/2017:18:00:05 -0700] "GET / HTTP/1.1" 301 562 "-" "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/49.0.2623.75 Safari/537.36 Google Favicon" 

I would consider that fairly dispositive. I just went into GSC and invoked the "your sites" dropdown. Those are two consecutive log lines from one site (it happens to be the one that just went HTTPS, hence the 301s). Another site had the same thing at the same time from 66.249.85.I-forget; the HTTPS site got 66.249.80.four-different-numbers; another got 66.249.80.some-other-numbers-again.

I have to say, tangentially, that it makes me a little uneasy that I'm seeing paired 200s instead of one 200 and one 301 at each place. I know I've got the appropriate redirects. Hm.

[keyplyr]

So both those are coming from Google Proxy, just different ranges. Google does use these ranges for some tasks (hence my allowing exceptions.) Never paid attention to the GSC requests though. However, Google also uses a small subset of proxy ranges to GET images posted at Google Plus.

Don't know what the double hits is about. Not sure if I've seen that before. If they are independent of each other, purposed for different reasons, that might account for the two 301s. They are from different ranges after all.