Zabbix wrong statistics

[ilik1989]

Hi i recently installed zabbix 2.0 on ubuntu 12.10 and have no item to show the main parameter in system like load average, also i have an item key vm.memory.size[available] which shows Available memory is defined as free+cached+buffers memory but all i need is only free memory.
Could anyone help me?

[tchjts1]

Hi i recently installed zabbix 2.0 on ubuntu 12.10 and have no item to show the main parameter in system like load average, also i have an item key vm.memory.size[available] which shows Available memory is defined as free+cached+buffers memory but all i need is only free memory.
Could anyone help me?

You can create your own item and key for free memory. Look at the item you have now, and at the bottom is a "clone" option. Clone it and change the information for the name of the item and make the key vm.memory.size[free].

See this page: https://www.zabbix.com/documentation...s/zabbix_agent

As far as your issue on Ubuntu... try updating to the latest 2.0.4 agent.

[ilik1989]

o thank you for advice of free memory.
1) i need load average more
2) And what is difference between zabbix_agentd 2.0.3 or 2.0.4?

[tchjts1]

2) And what is difference between zabbix_agentd 2.0.3 or 2.0.4?

Probably none. In your original post you mentioned 2.0 ... not 2.0.3

[ilik1989]

ok and what about the system main parameter load average which shows how far system is loaded?

[assie]

Load

ok and what about the system main parameter load average which shows how far system is loaded?

I am assuming here you think that the load graph is incorrect. Thats what I thought when I compared the load of 4 to a graphs showing 0.5.

This is actually, if you ready the graph carefully, correct. The machine has 8 cores, so the load PER core on 4 is actually 0.5 ( 4/8=0.5) So its a 'smart' graph, you dont have to know the number of cores to decide what a high load actually is

[ilik1989]

i assume you say about processor load, but i meant load average that can be seen by running top on bash. Load average on server is between 0 and 40

[assie]

i assume you say about processor load, but i meant load average that can be seen by running top on bash. Load average on server is between 0 and 40

Hm, not sure if we are miscommunicating here, whats your question here;

1) i need load average more

The processor load graph shows just that, the load, from e.g. uptime and/or top. Its divided like I said. Perhaps you mean something else?

Please ellaborate, then we are sure we're on the same page

[ilik1989]

The processor load graph shows just that, the load, from e.g. uptime and/or top. Its divided like I said

wikipedia

I meant load average that represents the average system load over a period of time

[assie]

wikipedia

I meant load average that represents the average system load over a period of time

Thats supported out-of-the-box, the zabbix agent supports

system.cpu.load[,avg1]

or 5/15..

Add that in a graph and you have the load.

Now we are on the same page, right?

[ilik1989]

Thats supported out-of-the-box, the zabbix agent supports

system.cpu.load[,avg1]

or 5/15..

Add that in a graph and you have the load.

Now we are on the same page, right?

that's sad cause the main parameter on any linux host is la.

[assie]

that's sad cause the main parameter on any linux host is la.

Its sad because zabbix supports it