agent.ping with nodata false triggers

[bashman]

Hi,

When I use this trigger:

{HOSTNAME:agent.ping.last(0)}=0 | {HOSTNAME:agent.ping.nodata(300)}=1

with this item:

Type ZABBIX agent
Key agent.ping

It false triggers for some hosts.

[bbrendon]

It seems like you're duplicating checks. Why not just pick one method to determine if an agent is down and then troubleshoot from there?

[bashman]

I did, but the nodata still false triggers.

I want to know when a host is dead, and the agent.ping.last doesn't alert me when a host is dead, that's why I want the nodata function.

Thanks for your response.

[MrKen]

In this post, Elvar says he fixed this problem by increasing the timeout in the zabbix_server.conf

http://www.zabbix.com/forum/showthread.php?t=15068

HTH

[bashman]

Hi Mrken,

I fixed my problem thanks to your post: http://www.zabbix.com/forum/showthre...6362#post56362

I checked if the hostname in the zabbix_agentd.conf was the same in the zabbix gui and It wasn't for the host that false triggered.

Thank you very much.

[bashman]

After a while the problem still remains, the zabbix_agentd log entry:

20683:20091229:104541 zabbix_agentd started. ZABBIX 1.6.6 (revision 7836).
20684:20091229:104541 zabbix_agentd collector started
20686:20091229:104541 zabbix_agentd listener started
20685:20091229:104541 zabbix_agentd listener started
20687:20091229:104541 zabbix_agentd listener started
20688:20091229:104541 zabbix_agentd listener started
20689:20091229:104541 zabbix_agentd listener started
20690:20091229:104541 zabbix_agentd active check started [<IP>:10051]
20686:20091229:113150 Got SIGPIPE. Where it came from???

Some items are not refreshing, the triggers for these items don't trigger due to Error:
Host is unavailable.

If I add to my trigger {HOSTNAME:agent.ping.last(0)}=0 this new condition:

{HOSTNAME:agent.ping.last(0)}=0 | {HOSTNAME:agent.ping.nodata(300)}=1

For the item:

Type zabbix agent
Key agent.ping
update interval 30
Keep history 7
Keep trends 365

My zabbix_agentd.conf:

ServerPort=10051
ListenPort=10050
StartAgents=5
#RefreshActiveChecks=120
#DisableActive=1
#EnableRemoteCommands=1
DebugLevel=3
PidFile=/var/run/zabbix-agent/zabbix_agentd.pid
LogFile=/var/log/zabbix-agent/zabbix_agentd.log
Timeout=3
UserParameter=oracle[*],/opt/zabbix/zabora $1 $2

It false triggers for some hosts. I had a switch problem the 22th, and some items don't refresh from that day.

I tried restarting the agents, checking the hostnames in the zabbix_agentd.conf are the same as in the zabbix gui (case sensitive), connection to IP.

related post: http://www.zabbix.com/forum/showthread.php?t=15112

[bashman]

It could be a timeout problem, I'll try to increase Timeout=3 to Timeout=5 in zabbix_server.conf.

[elvar]

For what it's worth, I posted a couple days ago on some agent.ping.nodata(300)}=1 threads saying that raising my timeout fixed the problem (which it did for one host) but after switching other hosts to that same trigger I kept having problems. Even if I upped my timeout to 30 (max) I still see this trigger fire even though the host is up and I see successful agent.ping attempts in the debug log output of the server.

23922:20091229:063129.511 In get_value_agent() host:'Host' addr:'Remote.IP.Here' key:'agent.ping'
 23936:20091229:063130.702 In evaluate_function('Host:agent.ping.nodata(300)')
 23936:20091229:063130.702 End of evaluate_function('Host:agent.ping.nodata(300)',value:'0'):SUCCEED

I have yet to hammer out exactly what's causing these false alarms but it's extremely frustrating. It feels like a bug but I can't pinpoint it. I did notice that the last time I got a value:'1' in the debug logs for that check it triggered the alert but I don't understand why it received a value of '1' when I can see the host is up, the agent is running, and all my manual ping requests are coming back fine. Is the remote agent bombing out? Is the Zabbix Server having issues? *sigh*

[bashman]

In the Zabbix Manual v1.6 I can see that agent.ping must have type "ZABBIX trapper". The host should periodically send data for this parameter using zabbix_sender.

[bashman]

I tried that, but I still get false alarms, my trigger agent.ping.nodata(180) doesn't work.

Can the function nodata work with agent.ping?

[MrKen]

In the Zabbix Manual v1.6 I can see that agent.ping must have type "ZABBIX trapper". The host should periodically send data for this parameter using zabbix_sender.

I must be going blind. What page is that on?

MrKen

[bashman]

Sorry it's on page 130 > Example 8:

Use of function nodata():
{zabbix.zabbix.com:tick.nodata(180)}=1
‘tick’ must have type ‘ZABBIX trapper’’. In order to make this trigger work, item
‘tick’ must be defined. The host should periodically send data for this parameter
using zabbix_sender. If no data is received within 180 seconds, the trigger value
becomes TRUE.

[bashman]

My trigger works fine now with this expresion:

{HOSTNAME:agent.ping.last(0)}=0 | {HOSTNAME:agent.ping.sum(#3)}<1

[bashman]

It works now with this expresion:

{HOSTNAME:agent.ping.max(300)}<1 | {HOSTNAME:agent.ping.nodata(300)}=1

zabbix agent version: 1.6.6
zabbix server version: 1.8

[1berto]

Having the same issue here, zabbix 4.0 :O ...