Is this weird or am I missing something ?

[soamz]

Hello, I just started Cacti and it seems solid, super powerful.
Works so well.

I have added my Microtik routers.
For example, one microtik ether is showing traffic of 150-179Mbps download and 30-45Mbps upload constantly since last 5 hours, but Cacti graphs says in bound 71Mbps and outbound 25Mbps.

Whats wrong ?

[micke2k]

Post screenshots

[soamz]

Post screenshots

Attached.

[soamz]

waiting

[micke2k]

Cacti graphs average by default. Could it be that mikrotik values are current values, or max values?

Cacti will graph the average value over time, not max values.

You can add max value manually, test that and see if you like the result better.

[soamz]

Cacti graphs average by default. Could it be that mikrotik values are current values, or max values?

Cacti will graph the average value over time, not max values.

You can add max value manually, test that and see if you like the result better.

All I need graphs to exactly reflect what is happening at microtik ethers.

And, I see its 1 min internal refresh.

Is there a way we can see everything in real time or even half a second ?

[micke2k]

You cant see it faster then you poll it. What is your poller setting? 5 min?

You can set poller to 10 sec, and change step and hearthbeat to 10 sec in the data source template. You need to delete and recreate it for it to work,

It will then show you 10 sec poller intervalls.


use prtg if you want for example 1 sec intervalls running on your screen.

Snmp isnt really designed for fast polling since it depends on the device to actually update the snmp. You can poll it 1000/sec but if the device doesnt update the snmp it doesnt matter.

[soamz]

You cant see it faster then you poll it. What is your poller setting? 5 min?

You can set poller to 10 sec, and change step and hearthbeat to 10 sec in the data source template. You need to delete and recreate it for it to work,

It will then show you 10 sec poller intervalls.


use prtg if you want for example 1 sec intervalls running on your screen.

Snmp isnt really designed for fast polling since it depends on the device to actually update the snmp. You can poll it 1000/sec but if the device doesnt update the snmp it doesnt matter.

Im afraid to touch the settings page now.
Last time, I touched, everything stopped working

Let me know, what exactly I should be doing, or I fear to touch the settings.

[micke2k]

You cant see it faster then you poll it. What is your poller setting? 5 min?

You can set poller to 10 sec, and change step and hearthbeat to 10 sec in the data source template. You need to delete and recreate it for it to work,

It will then show you 10 sec poller intervalls.


use prtg if you want for example 1 sec intervalls running on your screen.

Snmp isnt really designed for fast polling since it depends on the device to actually update the snmp. You can poll it 1000/sec but if the device doesnt update the snmp it doesnt matter.

Im afraid to touch the settings page now.
Last time, I touched, everything stopped working

Let me know, what exactly I should be doing, or I fear to touch the settings.

Yes you might have to recreate graphs etc for stuff to work with faster polling.

Maybe deploying a new cactiez in vmware will help? You can use it for labbing/testing and then implementing changes.

If you run it in vmware not just take a snapshot before you do anything and you can revert.

[soamz]

You cant see it faster then you poll it. What is your poller setting? 5 min?

You can set poller to 10 sec, and change step and hearthbeat to 10 sec in the data source template. You need to delete and recreate it for it to work,

It will then show you 10 sec poller intervalls.


use prtg if you want for example 1 sec intervalls running on your screen.

Snmp isnt really designed for fast polling since it depends on the device to actually update the snmp. You can poll it 1000/sec but if the device doesnt update the snmp it doesnt matter.

Im afraid to touch the settings page now.
Last time, I touched, everything stopped working

Let me know, what exactly I should be doing, or I fear to touch the settings.

Yes you might have to recreate graphs etc for stuff to work with faster polling.

Maybe deploying a new cactiez in vmware will help? You can use it for labbing/testing and then implementing changes.

If you run it in vmware not just take a snapshot before you do anything and you can revert.

Okay i will mess it up, so I pass