View only the new entries in a growing log file

It might suffice to use watch:

$ watch tail -n 15 mylogfile.txt

Instead of printing matching lines with decorations, decorate matching lines, and print everything:

awk '/User@Host:/ { $0 = "\033[32m" $0 "\033[39m" }; /Query_time:/ { $0 = "\033[29m" $0 "\033[39m" }; 1'

This awk program consists of three patterns and associated actions:

• lines matching User@Host: are processed with $0 = "\033[32m" $0 "\033[39m", which surrounds the current line with the given escape codes;
• lines matching Query_time: are processed with $0 = "\033[29m" $0 "\033[39m";
• lines for which the pattern 1 is non-zero (so all lines) are processed with the default action, which is to print the current line.

These are cumulative: all actions whose patterns match the current line are executed, in the order in which they are defined. A line matching User@Host: is modified (first action) and printed (third action). A line matching Query_time: is modified (second action) and printed (third action). A line matching neither is printed (third action).

That's the simplest technique to print all lines, without having to keep track of whether a line has been processed yet (to avoid printing a line twice). Instead of printing something in every set of action statements, we modify the current line as appropriate, and only print it once (but always print it).