Historical Graph Data (Inaccurate) Due to averaging?

Hello,

I was hoping someone could help me or point me in the right direction. I’m not a guru when it comes to snmp graphing :slight_smile:

We use LNMS to basically graph all our network devices. Firewalls, routers etc. Love this tool. However there is one slight issue I have run into. And I believe it is not limited to LibreNMS. I feel this is an issue I would run into using any other tool which relies on the RRDTool framework…

I am often asked to retrieve graphs for interface utilisation or concurrent session count by my team. When I focus on a graph in the last 6-12 hours, it is very accurate. I know this because, if we experience a major service interuption, which results in all IPSEC tunnels to drop on a device. If I focus on the window of impact the graph correctly shows the session count drop to 0.

However, let’s now pretend a month has elapsed and I want to go back to the exact window of said event. The min and max values are way different. Instead of zero it might be 100 or something much higher.

Why is this? Can I do something to change how this works? Any suggestions? I assume what is happening is, due to the “historical data” being averaged when i go back to focus on a specific time window the values are not precise anymore, but some kind of average.

Can I make it so that the last 6 months of data is 100% accurate? Disk space, etc wont me an issue. I’m sure my sys admin can hook me up :slight_smile:

Thanks for taking the time to read!