Why is utf8 encoding byte lost in php-fpm?

Question

PHP

There is a web server with Nginx and php-fpm on board.
usecase: We create a text file on the web server with content in Cyrillic in utf-8 encoding. If this file is given the .html extension, then Russian characters are normally displayed in the browser. If you rename it to .php, then Russian characters turn into "question questions" (????? ?? ????).
At the same time, if in the .php file the text is displayed not as a spool, but through <? echo "cyrillic"; ?> then the text is not "wrapped".
Tell me what could be the problem? Where can one byte of encoding be lost?

Answer 1

What encoding are the files in? I'm assuming it's not UTF-8, or UTF-8 BOM.
Convert files to UTF-8 without BOM, for example using Notepad++

Answer 2

F12 - Network - F5 - first line - Headers - look for the Content-Type header and wonder what is written there.