Recurring bug display on the traffic monitoring of interface

[J_C]

Hi,

I've been facing a recurring bug on some (not all) of the graph that I've implemented to monitor the traffic on Cisco router's interfaces.
This bug is recurring traffic gap and pic which occur very regularly (every 6min, or 8min). See pic attached to this post.

To make sure it's not a bug from the router, I've checked to monitor the same OID with a different tool (SNMPGrapher). And the problem is not there. There is no regular gap/pic occuring.

I monitor around 30 collectors from the same routers (mainly from OID related to the traffic load for each QOS class. I've the problem with some of the collectors but not with all of them.

I'm using Zabbix 2.0.8 on a CentOS 32bit.

Thank you very much in advance for your suggestion

[tchjts1]

I'm wondering if you have enough pollers configured to handle the workload.
Can you take a look at this post, the very last paragraph and the graphs that follow: https://www.zabbix.com/forum/showthread.php?t=41219

If you have that set up, it will give you a clear view of the performance of your Zabbix server and whether you need to make adjustments for your cache settings, pollers, etc.

Jeff

[J_C]

Hi Jeff,

Thank you very much for your answer. I've tried to gather all the information from Zabbix regarding the cache settings, pollers,...
It seems ok for me put I took screenshot and attached it to this post.

I still have the same problem and don't have any clue what could be the cause. This bug is really impacting my work
What's strange is the bug is not here 100% of the time. Graphs can be displayed for 1 hours without this problem of spike and gap.

For additional information, I'm using Zabbix 2.0.8 on a CentOS 6.3 x86 that I'm virtualizing with VirtualBox on a PC Latitude E5530 with a Proc Intel Core i7 3Ghz and 4GiG of RAM.
I've had a look to the server LOGS without seeing any obvious problems.

Do you think that upgrading Zabbix in version 2.2 could solve my problem?

Thanks a lot for your help.

JC