[ZBX-15406] Switch interface traffic graph is abnormal. Created: 2019 Jan 04  Updated: 2019 Jan 15  Resolved: 2019 Jan 09

Status: Closed
Project: ZABBIX BUGS AND ISSUES
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Problem report Priority: Major
Reporter: Claude Assignee: Zabbix Support Team
Resolution: Cannot Reproduce Votes: 0
Labels: graphs
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

centos 7.0, zabbix 3.4.14


Attachments: JPEG File actual values.jpg     PNG File image-2019-01-04-10-51-48-947.png     JPEG File item-conf.jpg     JPEG File preprocessing.jpg     JPEG File show values.jpg     JPEG File smaller period.jpg     JPEG File topologic.jpg     JPEG File traffic.jpg     JPEG File traffic1.jpg    

 Description   

The zabbix monitors the interface traffic of the switch. In the interface graph, there are two input peaks of more than 6Gbps, and the same traffic can be found from another interface.

However, this 6Gbps is incorrect. On the switch interface, the actual traffic peak is 1.4Gbps, and this is a 10GE interface.

The interface of the downlink switch also has the same abnormal traffic peak.

And the daily traffic never exceeds 1Gbps.

Use snmpv2c.

I wanna the graph to be correct. But I don't know what to do.

Please help.

 



 Comments   
Comment by Edgars Melveris [ 2019 Jan 04 ]

Hello Claude!

Can you please show us the configuration page for these items (also please include the preprocessing steps).

Also, can you select a smaller period in the latest data, which includes the spikes and switch to "Show values" and show us the actual values for these items

Comment by Claude [ 2019 Jan 04 ]

Hello Edgars:

I attached pictures. Please take a look.

Any questions let me know.

Comment by Edgars Melveris [ 2019 Jan 04 ]

When you look at the graph of latest data, please select "Values" at the "View as" menu. And then also add them here

 

Comment by Claude [ 2019 Jan 07 ]

I found and put screenshot named "show values" here.

Thanks.

Comment by Edgars Melveris [ 2019 Jan 07 ]

Well the graphs are showing correct data, you actually have data in the DB, that differs so much.
Zabbix does just simple math in this situation (change per second, multiply by 8) and just stores the value as is.
I would suggest looking at the device itself, to see why it returns such data. Does this happen frequently?

Comment by Claude [ 2019 Jan 08 ]

Thank you for your help.

But it happened on our two devices which connect directly. See attachment topologic. The traffic goes down. The five 10GE interfaces marked red got abnormal. GE interface was normal. Also eth-trunk interfaces got abnormal.

This is the only time abnormal. And traffic never exceeds 1G, except this time. Untill now our zabbix runs only 2 months.

Comment by Edgars Melveris [ 2019 Jan 08 ]

We can't be sure how the statistics are calculated on the device itself and how that differs from the values returned by snmp. My guess would be, that the device calculates some average for a time period, so even if that period is small, the numbers will differ, as in this case Zabbix saves an absolute change. So far, nothing indicates a bug on Zabbix side.

 But it happened on our two devices which connect directly.

I would believe that this also indicates, that there was a traffic spike between those devices.

Comment by Claude [ 2019 Jan 09 ]

I'm going to do some tests and get more info. Also study more about SNMP.
I'll post them here.
Thank you.

Comment by Edgars Melveris [ 2019 Jan 09 ]

Thanks for the update.
I'm going to close the issue for now, as right now there is no indication of a bug in Zabbix.
Feel free to reopen, if you find some more details.

Comment by Claude [ 2019 Jan 15 ]

I found the answer.
The switch calculates by 300 seconds(got this from manufacturer Huawei). So switch shows 1.4G.
Our zabbix calculates by 30 seconds. So zabbix shows 6G.
They are all correct. They should be different.

Comment by Edgars Melveris [ 2019 Jan 15 ]

Hi, thanks for the feedback, glad you've found the answer.

Generated at Sat May 09 17:57:37 EEST 2026 using Jira 10.3.18#10030018-sha1:5642e4ad348b6c2a83ebdba689d04763a2393cab.