How do I block a http user agent or a software agent using Nginx web server under Linux or Unix like operating systems?
You can block any http user agents with GET / POST requests that scrape your content or try to exploit software vulnerability. Use the following syntax. Edit /usr/local/nginx/conf/nginx.conf file, enter:
# vi /usr/local/nginx/conf/nginx.conf
In this example, block http user agent called wget:
## Block http user agent - wget ## if ($http_user_agent ~* (Wget) ) { return 403; } ## Block Software download user agents ## if ($http_user_agent ~* LWP::Simple|BBBike|wget) { return 403; } |
Save and close the file. Reload nginx web server, enter:
# service nginx reload
OR
# /usr/local/nginx/sbin/nginx -s reload
How do I block multiple http user agents?
Use the following syntax:
if ($http_user_agent ~ (agent1|agent2|Foo|Wget|Catall Spider|AcoiRobot) ) { return 403; } |
Case insensitive blocking: ~* vs ~
Please note the ~* makes it case insensitive as opposed to just a ~:
### case sensitive http user agent blocking ### if ($http_user_agent ~ (Catall Spider|AcoiRobot) ) { return 403; } ### case insensitive http user agent blocking ### if ($http_user_agent ~* (foo|bar) ) { return 403; } |
10 comment
How do you put another condition instead of return? can you deny all and only allow a certain ip/subnet with those user agents to go through?
How do you put another condition instead of return? can you deny all and only allow a certain ip/subnet with those user agents to go through? any example?
i get the following error : “nginx: [emerg] “if” directive is not allowed here in /etc/nginx/nginx.conf:”
any thoughts?
You need to make sure your if statement is in the server block.
Thanks Nigel, that did the trick.
its working perfectly now.
What is server block meaning ?
I got the same problem nginx: [emerg] “if†directive is not allowed here
Question, is how to block a block that spoofs its user-agent, where it’s clear that the user agent is spoofed by the Comments token.
i.e. the following user-agent string:
“Mozilla/5.0 (compatible; ACHE/Unknown Version; +https://github.com/ViDA-NYU/ache; )”
Where we don’t want to be blocking by Mozilla/5.0, but rather want to be blocking by the entire string. (i.e. how to make the instruction look for “ACHE/Unknown Version” within the comments token of the user string.)
Thats the real question, what’s the point of blocking useragent when every bot uses Mozilla/5.0, and after that some random shi.t
Even we block http agent here (in my case its Jorgee) ,
how can we deny it, even without returning 403 for the request.
403 is denying it. It won’t able to do anything else on your server.