1

I have been tasked with finding the cause of sporadic performance issues on a website that I didn't build or host, after doing a number of tests I suspect it may be an issue with the network rather than the server or the webapp its self.

I was wondering if anyone knows of a good too do monitor network performance remotely over a period of time, for example I would like to do a traceroute repeatedly over several hours and record the results.

1
  • 1
    traceroute really isn't the right tool for performance monitoring\measuring.
    – joeqwerty
    Nov 22, 2010 at 13:36

3 Answers 3

1

From your description it sounds to me like you are looking to do a quick one-off solution to this monitoring situation that something permanent and robust. If this is true than I recommend seeing if you can:

  1. Quickly set up pingdom to get latency as an hourly average. If you need more detail, just leave a ping running from your home PC.
  2. Contact that ISP to see if they keep network graphs.

If you want to do something more extensive than I would recommend both a remote and local Nagios installation to monitor all of this. You will also be able to capture the hosts resource usage.

Lastly, for what it is worth I usually find it is not the network if you have decent hosting. Make sure you checked all the logs. scheduled tasks, and monitored the DB for long running queries.

2
  • ISP in question is a 3rd tier ISP and the performance of the website in question is worse and has more sporadic connection problems in the same building than on my home broadband, every test I have run suggests the server, the webapp and the database are all humming along nicely. But if i go back to the network guys and tell them I think its there problem they won't listen unless I have evidence.
    – 3urdoch
    Nov 22, 2010 at 13:50
  • In that case I would just leave a ping running from the building. Once that shows packet loss or high latency that should be enough evidence ... Nov 22, 2010 at 13:55
1

A tool like MTR (or WinMTR) or PingPlotter that tracks per-hop (like traceroute) latency over time can be really useful for a problem like that - it can allow you to see where in the network path the problem is.

0

Check out SmokePing. It does pretty-much exactly what you are asking.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .