Get rid of requests library

[krizhanovsky]

Since it's not trivial to get response status code from Selenium, at the moment we use requests to get the status. This increases requests rate and reveals the robotic nature of the bot. We need to do everything with Selenium only. To achieve this we can parse the page body and/or get (intercept) data from the socket layer.

[krizhanovsky]

One more thing I discovered in the current usage of requests is that it uses HTTP/1. It can not fetch any of the resources from our website, e.g.

WARNING:root:BROKEN LINK https://tempesta-tech.com/blog/ on https://tempesta-tech.com/ (can not connect)

and there is

28735.672417] [tempesta tls] Warning: [::ffff:31.221.30.162] ClientHello: cannot find matching ALPN for http/1.1

on Tempesta FW side.