Note that the agent would require super-user privileges to do that.
Maybe we could use the same approach as with the server, fping installed with setuid?
Sooo, any progress on this? 
Going the fping way would be great.
I have a workaround for this.
UserParameter=chk.fping[*],fping -c 3 $1 2>&1 | tail -n 1 | awk '{print $$8}' | cut -d/ -f2
But the fact that stuff like this needs to be worked around means these guys need more help and community code support as they cannot implement new features fast enough. This should not be a 3 year old ticket. Nor is it a minor feature 
. I don't know if it's a devops or an agile problem. But zabbix has a small window to make some major new things happen to satisfy the changing marketplace of SaaS and application monitoring. A UI overhaul, better (MODERN) graphing, and general expansion on their greatest strength in expressions.
I'm considering forking the code and making some changes in my personal feature request queue. In particular in the UI and in expression handling and item storage flexibility, item and string comparison evaluations, easier construction and better testing of triggers, inventory support, using zabbix as a ENC for puppet / chef, Server-Agent SSL, etc.
Jira is messing up my UserParameter line.... it's all one line, and no spaces in between the single quotes in awk.
tagwolf, instead to go for a fork may I encourage you to collaborate by providing small patches?
If so, you should take a look at the coding guidelines to assure your patches could be merged without much editing.
Btw, you might possibly want to add youre one-liner here. Another awesome event to collaborate are Bugsquash days.
For further arrangements or discussions look in at IRC or forums.
I'd love to do some patches according and have no problem following your coding standards.
If though I have a larger modification, who would I talk to about user stories, testing and possible integration with the main codebase?
Also I think it's awesome you guys are doing bugsquash days! I'll check them out.
it's probably best to start with really small patches and push those through the bugsquash days, otherwise large patches are unlikely to get any attention (unfortunately) - but the best place for such discussions is irc
Hi
I improved the Userparameter a bit for those of you who want to know:
With this slight changes you get a 0 value from fping output in case the remote host is not reachable, therefore making it easier for a trigger to be defined (0 value means unreachable).
Instead of defining a fixed column ($$8) I use the last column with $NF. and as the middle value of every output is either avg latency, or received packets, it will always be a value can measure stuff.
UserParameter=chk.fping[*],/usr/bin/fping -c 3 $1 2>&1 | tail -n 1 | awk '{print $NF}' | cut -d '/' -f2
PS: on my FreeBSD machine i had to provide the full path, so please (to all) check the correct path to fping ('type fling') and always provide it.
PPS: But still, a built in function would be the best!
Best regards
Daywalker, you might want to take a look at User parameter one-liners too.
Looking for simple way to mesure ping response times from host to host, for example from remote machine to zabbix server. To monitor situations where ping goes >100 ms for example.
icmppingsec do the job in one direction, but I need in reverse direction, too.
Yes, ping. Of course. Absolutely needed. Very basic. Low hanging fruit. Please do it already. I have never seen any network monitoring setup that does not use ping - it is the first thing every experienced admin will implement across networks / host groups / etc. So with Zabbix you have currently this experience: ok, lets implement ping for all the things - oh, Zabbix does not support that out of the box, I need to fiddle around with manually editing config parameters and such, o my god.
Please implement that. Thanks. Or at least implement something similar that zabix-agent will do out of the box by default.
This request can be implemented as a loadable module (i.e. outside of Zabbix code base) if there is enough interest. ZBXNEXT-2735 mentions liboping, module can simply wrap it. Or to avoid granting capabilities to Zabbix binaries there can be a standalone daemon to do the pinging and the module will simply query that daemon.
This feature is indispensable, for example, to check if a remote server is reachable from the intended client (not from zabbix server).
In the meanwhile, here's a similar workaround for Windows agents:
UserParameter=ping.host[*],ping -n 1 $1 > nul && echo 1 || echo 0
For Unix you can use something as simple as
UserParameter=ping.host[*],/usr/bin/fping -q -r 3 $1 ; echo $?
Eliminates the need for screen-scraping fping output, "0" means host is reachable and "1" means not reachable.
But I agree it's high time to implement this in the agent.
FYI there is a patch for agent2 from one of community members, klimenko.andrei:
It would be useful if we could use icmpping like we use net.tcp.service...
@Nicola Mauri Hello.
So i tried your UserParameter for Windows but it didnt work... ping -n 1 192.168.1.1 > nul && echo 1 || echo 0 always returns 1 even when its disconnected.
Yes I would also really like to see this implemented.
I'm also struggling with getting the userparameter working successfully, works when connected but when timeout occurs item goes to Not Supported with info Timeout while executing shell script. Agent log contains below, and of course the trigger won't work.
7760:20211207:070606.891 Failed to execute command "ping 10.10.0.11 -n 1 |find "TTL=">NUL && echo 1 || echo 0": Timeout while executing a shell script.
4116:20211207:070608.024 Failed to execute command "ping 10.10.0.12 -n 1 |find "TTL=">NUL && echo 1 || echo 0": Timeout while executing a shell script.
Try to set the ping timeout lower than the agent/server timeout - if the ping timeout is longer, the agent/server communication times out first.
This simple common function should not be so difficult and inconsistent to implement. The agent should be able to run ping just like the server simple check.
another way, native : wmi.getallroot\cimv2,"SELECT PrimaryAddressResolutionStatus, StatusCode, ResponseTime FROM Win32_PingStatus WHERE Address='{#ADDR}' AND Timeout=2000"
Maybe this will help someone.
icmppingsec for Zabbix Agent on Windows.
UserParameter=icmppingsec[*], powershell -NoProfile -ExecutionPolicy bypass -Command "Test-Connection -ComputerName $1 -Count 4 | Measure-Object ResponseTime -Average | ForEach-Object { '{0:N5}' -f ($_.Average / 1000) -replace ',', '.'}"