Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials
Store

Configure max bandwidth on interface

Hello all,

I have a question that should be simple but apparently isn't...

I'm attempting to monitor bandwidth utilization at remote sites using NPM, however I keep getting incorrect information. A number of our smaller remote sites only have a single T-1 circuit, but SolarWinds still treats them as 100MB since the monitored interface has a throughput of 100Mb. I can't monitor the actual device that has the T-1 circuit directly connected (long story), so I want to get accurate data from the closest interface to that device (i.e. another router or switch). As a test, I manually set one of those interfaces to a custom Bandwidth of 1.544 Mb/s (Transmit and Receive) but it still doesn't show correctly on my network utilization map, or when I run performance metric tests.

I need to collect this information in order to justify bandwidth increases at most of my remote sites, so I need the data to be as accurate as possible. I do also have NetFlow running on my routers and switches that support it, that is sending data to SolarWinds as well.

Find more posts tagged with

Accepted answers

mesverrum

Did you force that device to poll after setting the custom bandwidth? They don't factor those changes in until the next poll.

All comments

mesverrum

Did you force that device to poll after setting the custom bandwidth? They don't factor those changes in until the next poll.

tom.diehm

Looks like that was part of it, I had to manually force a poll. I've had it set for a few days, so I assumed that it would have automatically polled by now, but I guess not.

That said, the numbers on PerfStack still don't seem to be correct when I run a file copy test. A single T-1 should be 1.544 Mbps, but I'm seeing a max of 1.97 Mbps on my switch during the test, and a percentage of only 2% (although the percentage seems to be closer to correct on my map now, reading 59%).

mesverrum

What is your interface statistic polling interval and what is the time range set on the chart?

SNMP bps calculations are actually an average, they poll the counters for total bits sent/received and then divide that by the amount of time between the current and previous polling interval, so it never really accurately reflects bursts in bandwidth, just the average between to timestamps. With that said I would expect normally it should read lower than your actual cap, not higher. Might need to get a few polls in a row in for it to get settled in, or maybe forcing the poll skewed the way it was doing the averaging calculation.

If it were me I'd set my chart to only show data since I made the change and then collect a few hours of data to get a sense of how closely the numbers are tracking with my expectations or if something still seems way off.

tom.diehm

I think I've got this working now, I manually set the internet-facing interfaces (at least, the closest to the internet that I'm able to monitor) to the circuit's bandwidth, and I appear to be getting correct information now. My best guess is that some of the graphs still incorrectly report the data, since the port itself might be 100 but the circuit is only 1.544.

I'm going to continue monitoring for a few days but it looks like I'm getting enough correct data from some of my graphs that this will work.

Thanks!

rschroeder

I've had success manually setting the L3 devices' uplink bandwidth. Some of my sites have used bonded T1's, others used small broadband uplinks of 5 Mbs. Manually editing each router's uplink port bandwidth and setting it to the actual amount allowed/contracted/policed has worked well for displaying the percentage of bandwidth used, no matter whether the port's physical capabilities were 100 Mb or Gig.

Force a new discovery of the device, then do three or four manual polls of it. That's been the ticket for me to get the display to reflect the actual amount of contracted bandwidth allowed.