[smokeping-users] Graph throughput distribution via SNMP

Patrick, Robert (CONTR) Robert.Patrick at hq.doe.gov
Mon Jul 14 17:55:05 CEST 2014

Assuming the request is track bandwidth utilization at a specific point in the network, such as the WAN uplink on a branch office router, where a Smokeping instance at an upstream data center is pinging that branch router over the WAN.  Yes?

Bandwidth tracking tools are available for side-by-side comparison with a Smokeping graph, but the tools don't always have graphs that line up at the same timeframes and steps that are used in Smokeping, and how to best align the vertical axis when trying to show bandwidth I/O versus RTT?

Rendering the current bandwidth utilization inside the same graph (or overlay) tracking with Smokeping response times would be useful, as increasing bandwidth utilization tends to drive increasing RTT.

In an example graph, four lines plus smoke -- red shows inbound bits per second at router X interface, blue shows outbound bits per second at router X interface, green shows RTT average to router X, and gray smoke for RTT jitter.

One option is that Smokeping polls the device directly for a specific device interface input and output bits, another option is for Smokeping to read the values from an RRD file or alternate local input method.

Yet another method is to have multiple polling tools stuffing data into RRD files (with matching steps, etc.) and then use a graph rendering web interface (e.g. DDRAW or equivalent) to consolidate the view.  This would allow combining multiple metrics into one graph, such as CPU load at one or more hops in the path, I/O load at one or more hops in the path, etc.

-----Original Message-----
From: smokeping-users-bounces+robert.patrick=hq.doe.gov at lists.oetiker.ch [mailto:smokeping-users-bounces+robert.patrick=hq.doe.gov at lists.oetiker.ch] On Behalf Of Niall O'Reilly
Sent: Monday, July 14, 2014 6:12 AM
To: Greg Sloop
Cc: smokeping-users at lists.oetiker.ch; Distant Smokes JPP
Subject: Re: [smokeping-users] Graph throughput distribution via SNMP

At Sun, 13 Jul 2014 23:38:29 -0700,
Gregory Sloop wrote:
> Nagios may be able to query SNMP, but it won't keep long-term stats 
> like MRTG or Smokeping do.

  Besides, MRTG can be configured to detect threshold-crossing events
  and run a relevant script.  This could signal the event to Nagios.

  Best regards,
  Niall O'Reilly

smokeping-users mailing list
smokeping-users at lists.oetiker.ch

More information about the smokeping-users mailing list