False packet lost alert from Nagios

[imran_khan]

Hello,

Thanks for the reply.

As per my knowledge, 40% warning and 80% critical is very high. Please suggest me practical value for these. Currently we are using below command to monitor ping of servers.

check_command check_ping!30000,1%!100000,2%

Thanks,
Imrna Khan.

[tmcdonald]

The lowest usable values you could have would be 20% and 40% because you are only sending out 5 packets and you cannot possibly have 10% packet loss with 5 packets. The math just doesn't work that way. You can't have 1% packet loss and you can't have 2% packet loss for the same reason. The "practical" or "acceptable" value is yours to decide.

[eloyd]

First, you need to figure out how many packets to send (5 or 10 would be good numbers). Then, you need to figure out how many packets of the 5 or ten you sent would be considered okay to lose. If you send five, and you consider 1 packet okay to lose, that's 20%. If you send 10 and consider one packet okay to lose, that's 10%.

If you send five packets (the default), each packet lost is a 20% overall loss (one out of five is 20%). If you send 10 packets, each packet lost is 10% overall loss (one out of ten is 10%).

For comparison, we use 60% for WARNING and 90% for CRITICAL and we send 10 packets.

[imran_khan]

Hello,

As per check_ping plugin default number of packet send is 5 after every 10 second.

-p, --packets=INTEGER
number of ICMP ECHO packets to send (Default: 5)

-t, --timeout=INTEGER
Seconds before connection times out (default: 10)

Currently warning value is 1% and critical is 2% but why we are getting Packet loss 16%, 28%, 40% etc?
If 1 packet loss out of 5 then packet loss should be 20%, for 2 packet loss it is 40%.

e.g.:- PING CRITICAL - Packet loss = 16%, RTA = 2.98 ms

Thanks,
Imran Khan.

[imran_khan]

Hello,

Here is more example.

PING CRITICAL - Packet loss = 28%, RTA = 2.69 ms
PING CRITICAL - Packet loss = 16%, RTA = 1.46 ms
PING CRITICAL - Packet loss = 16%, RTA = 1.29 ms

Thanks,
Imran Khan.

[imran_khan]

Hello,

Any help? Any update?

Thanks,
Imran Khan.

[tmcdonald]

What is the *exact* command you are running to test this? According to your definition above you have something like this:

Code: Select all

check_ping -H 192.168.5.41 -w 30000,1% -c 100000,2% -p 5

but are you testing with anything different from the command line?

And just as a heads-up, our offices are closed on weekends so unless someone from the community responds you will likely not be hearing from anyone at Nagios until Monday mornings.

[imran_khan]

Hello,

Defination of command is as below but in the service we did not define number of packet. By default check_ping consider packet 5.

define command{
command_name check_ping
command_line $USER1$/check_ping -H $HOSTADDRESS$ -w $ARG1$ -c $ARG2$ -p 5
}

Service:-
check_command check_ping!30000,1%!100000,2%

Please suggest me why we are getting below type critical alerts in nagios?

PING CRITICAL - Packet loss = 16%, RTA = 1.29 ms

Thanks,
Imran Khan.

[slansing]

Well, the simplest explanation is because you are set to swap to a critical state when you have >2% packet loss. You are currently at 16%, which would of course mean you would be in a critical state.

[imran_khan]

Hello,

I am still confuse about warning and critical value.

By default check_ping command send 5 packet in 10 second so what will be the warning and critical limit for this?
Please suggest me the value and correct me if I am wrong.

-p, --packets=INTEGER
number of ICMP ECHO packets to send (Default: 5)

-t, --timeout=INTEGER
Seconds before connection times out (default: 10)

Thanks,
Imran Khan.