Nagios Monitoring Data Consumption

[oteroj92]

Hello,

Our setup:
We have setup Nagios XI through Amazon's Web Services (AWS) in an instance.

I recently configured SNMP monitoring for multiple remote devices that have access to the internet through the cellular network. They have been set up for over a month. The issue that recently came up was the amount of data that was being consumed by these sites. One of them was as high as 60 GB for the month of February alone. I have attempted to eliminate Nagios from the list of possible causes by limiting what OID's I monitor (~5 per site) and how frequent I do polling (~5-10 minutes).

My question is: Have you seen any instances where Nagios requests massive amounts of SNMP data? if so is it due to bad configuration or is it more of a hard coded issue?

If you have never seen this issue, do you think Nagios could be the cause of high data usage for SNMP monitoring?

Thanks

[tmcdonald]

A few thoughts:

• Yes, this could be a misconfiguration. If you ran the snmpwalk wizard and left everything checked that would certainly explain the large amount of SNMP checks.
• MRTG is a backend component we use in our Network Switch / Router Wizard and it relies on SNMP exclusively for its checks. If you have run that wizard, there would be another source of the traffic.
• Depending on the number of sites you are checking, 5 per site could add up to a lot of checks. This might just be normal behavior.
• There is always the possibility that there are other things installed on the server that are adding to the traffic, but it's not likely anything we would install by default.
• Do also note that if a check fails it tends to start running a lot more frequently since there is a problem that needs to be watched more closely, so if something major happened in February which caused your checks to fail, you're potentially looking at 5x to 10x as much traffic during that period (5-10 minutes for a normal check, sped up to 1 check per minute).

[oteroj92]

Just to be clear, when you mean by "left everything checked" you mean the OIDS that I am trying to monitor from the device right?

I made sure to check only the services I needed, and not every single thing that the device was offering. However on some sites I did not change any part of the configuration when I ran the wizard (left at one check every five minutes, if critical check every minute for 5 minutes, etc)

[dwhitfield]

Did you open a ticket about this? We had someone email with a very similar issue. It's best to have a single issue in one place.