Page 3 of 5
Re: Latency Uptime Warnings/Alerts and Juniper Switches
Posted: Fri Dec 08, 2017 9:54 am
by emssa
So we decided to go for the tcping approach and edited the http check command as suggested via the pdf.
So we changed
$USER1$/check_http -H $HOSTADDRESS$
to
$USER1$/tcping $HOSTADDRESS$ 80
and sent it to all HTTP services for each host.
We now see:
Screenshot from 2017-12-08 09-49-18.png
Re: Latency Uptime Warnings/Alerts and Juniper Switches
Posted: Fri Dec 08, 2017 11:26 am
by kyang
Are you using check_tcp? for tcping? Or which plugin?
You could also try doing it through check_http, but I'll post the difference.
Here is my examples to show the difference.
check_http
Code: Select all
[root@localhost libexec]# ./check_http -H 192.168.4.125 80
connect to address 80 and port 80: Invalid argument
HTTP CRITICAL - Unable to open TCP socket
[root@localhost libexec]# ./check_http -H 192.168.4.125 -p 80
HTTP OK: HTTP/1.1 200 OK - 3271 bytes in 0.001 second response time |time=0.001040s;;;0.000000 size=3271B;;;0
Code: Select all
[root@localhost libexec]# ./check_tcp -H 192.168.4.125 80
connect to address 192.168.4.125 and port 0: Connection refused
[root@localhost libexec]# ./check_tcp -H 192.168.4.125 -p 80
TCP OK - 0.000 second response time on 192.168.4.125 port 80|time=0.000111s;;;0.000000;10.000000
I hope this helps!
Re: Latency Uptime Warnings/Alerts and Juniper Switches
Posted: Sun Dec 10, 2017 8:50 am
by emssa
Did it on all three and on check_ping but that seems to break.
It wants or is expecting an integer or percentage, it then complains about how it cant handle %s.
So maybe that one can not be changed? Is there a list somewhere with the ones that can be changed to tcping?
Also where is the https 443 check being called from as I do not see it in the drop down list of commands?
Re: Latency Uptime Warnings/Alerts and Juniper Switches
Posted: Mon Dec 11, 2017 9:55 am
by emssa
ah according to the very small readme the check tcping only supports 80, 22 and 23. So not an all around fix for host ping, 443 etc..
So you cant completely get away from ICMP checks?
Re: Latency Uptime Warnings/Alerts and Juniper Switches
Posted: Mon Dec 11, 2017 10:50 am
by kyang
I also looked at the readme and saw it.
You can't ping a specific port, but you could try using a telnet plugin to check if you can connect to a port? If you didn't want a ICMP check.
Code: Select all
[root@localhost libexec]# ./check_telnet.py -H 192.168.4.125 -p 5000
CRITICAL: Telnet connection failed
[root@localhost libexec]# ./check_telnet.py -H 192.168.4.125 -p 443
OK: Telnet up
Re: Latency Uptime Warnings/Alerts and Juniper Switches
Posted: Tue Dec 12, 2017 11:54 am
by emssa
Is there a step through on how to move completely off of ICMP then to follow?
Re: Latency Uptime Warnings/Alerts and Juniper Switches
Posted: Tue Dec 12, 2017 2:06 pm
by kyang
I'm not exactly sure what you mean.
You want to use ICMP, but then move off to follow?
Re: Latency Uptime Warnings/Alerts and Juniper Switches
Posted: Tue Dec 12, 2017 2:18 pm
by emssa
No we would want to move off ICMP to tcping.
Is there an appropriate walk-through or set of steps to do that?
Re: Latency Uptime Warnings/Alerts and Juniper Switches
Posted: Tue Dec 12, 2017 4:05 pm
by kyang
You could create a new template and then assign that plugin and check_command to be used with that template.
That's a pretty solid way of doing it, and that way you aren't editing the original templates that are there.
Along with that, if you have an enterprise edition of XI. You could use the bulk modifications tool to pass that template to all hosts if you wanted to.
Let us know if that helps!
Re: Latency Uptime Warnings/Alerts and Juniper Switches
Posted: Wed Dec 13, 2017 9:04 am
by emssa
So I will take that as a no.
So lets do this step-by-step I guess...
1. Go to Core Config Manager -> Commands (left Nav column)
2. Select Add New (Blue Icon Button)
3. Enter Information to look like this:
Screenshot from 2017-12-13 08-20-26.png
Is the above correct or should the command line be something else?
If all good what is the next step?