i believe most search engines only obey robots.txt at the root level, and many zabbix installations would use sublevel like /zabbix...
thoughts on that ?

They indeed take the toplevel one, but I don't know whether people use it as subdomain or with /zabbix. Generally people have a separate server (or virtual machine) with Zabbix. We use nginx+fpm which is a whole lot better performance wise and have configured the webroot to use /usr/share/zabbix. Its just a simple static two-line file, even if it doesn't help everyone there would still be people who benefit from this

suggested contents of the example robots.txt.
any comments ?

# If Zabbix frontend is available on the internet, it is suggested to disallow
# access to it for the robots (like search engine crawlers), otherwise they
# may overload Zabbix system without offering any benefit.
#
# Note that this should not be used for security reasons, as any visitor is
# free to ignore contents of this file (or use them to know which directories
# are supposed to be hidden).
#
# The example configuration below will disallow any robot access if Zabbix
# frontend is in the root directory.

User-agent: *
Disallow: /

# If Zabbix frontend is in a subdirectory - for example, 'zabbix' - this file
# should be placed in the root directory of web documents directory and path to
# subdirectory should be specified like this instead:
# Disallow: /zabbix/

Looks good to me!

example robots.txt added in svn://svn.zabbix.com/branches/dev/ZBX-7811

(1) I've found the text a bit hard to read, so I've edited it a bit. Please review.

# If the Zabbix frontend is available on the internet, it is suggested to disallow
# access to it for robots like search engine crawlers. Otherwise they
# may overload the Zabbix system without offering any benefit.
#
# Note that this must not be used for security reasons, as any visitor is
# free to ignore the contents of this file or use it to know which directories
# are supposed to be hidden.
#
# The following configuration should be used if the Zabbix frontend is installed
# in the root directory of the web server. It will deny access to the Zabbox frontend 
# for all robots.

User-agent: *
Disallow: /

# If the Zabbix frontend is installed in a subdirectory, for example "zabbix", the
# "Disallow" directive must be changed to point to that directory: "Disallow: /zabbix/".

# It is important to make sure that the "Disallow" directory points specificly to the
# Zabbix frontend directory, since otherwise it may interfere with other websites
# running on the same domain.

Two important notes:

• I've written "Disallow: /zabbix/" inline, as part of a text, since users might accidentally uncomment the line and forget to comment "Disallow: /" leaving two directives in effect.
• I've added a warning for users to make sure that robots.txt only affects the zabbix frontend, not other websites on the same domain.

Misusing the robots.txt file may cause big problems for other websites on the same domain, so I think it's important to highlight these issues.

martins-v Two typos need to be fixed - Zabbox -> Zabbix, specificly -> specifically.

I think we may also omit the definite article in "the Zabbix frontend", since we perceive it as a unique proper noun, we are not singling it out against any other frontends, because for us we have only one. Then again, proper nouns should be capitalized, thus "Zabbix Frontend" would be totally correct, which again looks weird because it's not a brand for us or such. So, probably the best user friendly solution is to have "Zabbix frontend" without the definite article. Note that "frontend" without "Zabbix" should always be used with the definite article like "in the frontend".

<richlv> updated based on that, also reformatted to fit 80 char line length (assuming that this file might be more likely to be viewed from various terminals)

# If Zabbix frontend is available on the internet, it is suggested to disallow
# access to it for robots like search engine crawlers. Otherwise they may
# overload the Zabbix system without offering any benefit.
#
# Note that this must not be used for security reasons, as any visitor is
# free to ignore the contents of this file or use it to know which directories
# are supposed to be hidden.
#
# The following configuration should be used if Zabbix frontend is installed
# in the root directory of the web server. It will deny access to Zabbix
# frontend for all robots.

User-agent: *
Disallow: /

# If Zabbix frontend is installed in a subdirectory, for example "zabbix", the
# "Disallow" directive must be changed to point to that directory:
# "Disallow: /zabbix/".

# It is important to make sure that the "Disallow" directory points specifically
# to the Zabbix frontend directory, since otherwise it may interfere with other
# websites running on the same domain.

martins-v Seems good enough to me.

jelisejev Good, CLOSED.

TESTED.

(2) listed in :

• https://www.zabbix.com/documentation/2.0/manual/introduction/whatsnew2012
• https://www.zabbix.com/documentation/2.2/manual/introduction/whatsnew224
  please review

martins-v Reviewed, seems good. CLOSED.

merged to 2.0 in r44589, 2.2 in r44590 and trunk in r44591.