The bot for link previewing in Telegram currently uses a User-Agent string of TelegramBot (like TwitterBot)
. From my observation, it seems Cloudflare Analytics currently classifies it as TwitterBot, which is reasonable, but will cause confusion to users. Given Telegram’s popularity in some regions, it should make sense to distinguish Telegram traffic from others.
Thanks for bringing this up! Could you share more information about the potential TelegramBot
? Sharing the IP and UserAgent would be a good start. From there we can try to verify if the bot is indeed a TelegramBot and then check to make sure that we properly classify this traffic.
The User-Agent string it uses is just TelegramBot (like TwitterBot)
. When a user types a link into the popular messaging software Telegram, it will attempt to generate a summary of the linked page. This feature is called “link preview”.
The crawler in their server uses the aforementioned UA string. They dont’t publish the range of IPs they use, but I guess these crawling requests probably originate from their ASNs. A (maybe partial) list of those ASN can be seen in bgp.he.net.