agent.ping notification

[clayton940]

Hello All,
A relatively new user using Zabbix Appliance 2.0.4 .

I have been tuning zabbix and like the flow but I have seen a problem with the agent.ping notification.

I had a host go down that had the zabbix agent installed and I received a confusing message

Trigger: Zabbix agent on foo is unreachable for 5 minutes

Trigger severity: High

Trigger status: PROBLEM

Agent ping: Up (1)

My problem comes into is the Agent Ping status as I have modified the nofication to show the {TRIGGER.VALUE1}: {ITEM.VALUE1}

And it relates that the agent is up when in fact the agent is down. Do I have something configured wrong?

Here is the trigger for the agent.ping

{Template OS HP-UX:agent.ping.nodata(5m)}=1


And here is the notification action

Trigger: {TRIGGER.NAME}

Trigger severity: {TRIGGER.SEVERITY}

Trigger status: {TRIGGER.STATUS}

{ITEM.NAME1}: {ITEM.VALUE1}

Thanks for any help offered...

[filipp.sudanov]

The most resent version of Zabbix Appliance is 2.2, may be you could try it.

May be you could provide screenshots of trigger configuration - how it was before, when it was working and how it is after modification, when it's not working as expected.

[clayton940]

The trigger is the default I haven't modified it.

As for the upgrade to 2.2 appliance I would do that if there was an upgrade procedure but there isn't and since the 2.0.4 appliance is built on SuSE from the binaries/repositories and the repositories haven't been updated it makes it a tad bit rough.

But attached are the screen shots...

[filipp.sudanov]

I don't see anyting wroing in the screenshots.

As for the upgrade - I mean that you could just download the fresh image of the appliance. You can try to copy the configuration from one appliance to another as described here: https://www.zabbix.com/documentation..._export_import. There is no way to copy history - anyway the appliance is intended for testing, not for serious production, as the download page says.

[clayton940]

What does your agent ping notification look like? Are you using an agent?

[filipp.sudanov]

I've tried to create an action with the following in the message section:

1. {ITEM.NAME1} ({HOST.NAME1}:{ITEM.KEY1}): {ITEM.VALUE1}

and when the trigger goes into PROBLEM state, the message gots resolved into

1. Agent ping (192.168.1.13:agent.ping): Up (1)

So now I probably understand what you meant - that it's saying "Up" instead of saying "Down". Sorry, that I did no get you from the very beginning.

The thing is, that {ITEM.VALUE1} just takes the _last_ value stored in the database. But agent.ping only stores values of 1 into the database, or does not store anything when agent is not available. It does not store 0 into the database. That's why nodata(30s) is used in the trigger, not something like last().
So this behavour is pretty much by design and {ITEM.VALUE1} can not be used in action's message in that case.

[filipp.sudanov]

UPD:

Well, there actually is another item, that is simular to agent.ping, but stores a value every time, even when agent is down.

zabbix[host,agent,available], please note that it should be of type "Zabbix internal".

https://www.zabbix.com/documentation/2.2/manual/config/items/itemtypes/internal

[clayton940]

Nice find. I'll have to try it today...