Page 1 of 1

what am I seeing

Posted: Mon May 17, 2010 2:11 pm
by wisperer
when I scroll over too one device, I see the graph showing ping, http and ssh

I understand the ping time result but is the http showing traffic?

Why the ssh, would the result not be much the same as a ping?

thanks

Re: what am I seeing

Posted: Mon May 17, 2010 3:35 pm
by lebowski
When you auto discover devices services get added automatically. Typically you would want to setup the dude to only install services that you are interested in.

Ping gets added to almost everything since almost every device responds to ping.

SSH was added to that one system since the dude detected that system supports ssh.
HTTP was added for the same reason.

The graph for SSH and HTTP are round trip times in milliseconds(ms), how fast the server responds to probes.
You should see a small ms in the description of the graph. (since all 3 should be ms they should all be on the same graph)

They all do the same thing, test to see if the service is up.

The only time you will see traffic is on a link (unless you build a custom probe). Traffic on a link is all traffic and can't be viewed as its component parts.

HTH,
Lebowski

Re: what am I seeing

Posted: Mon May 17, 2010 3:35 pm
by lebowski
Oh Ignore the RTT of ping, it is woefully inaccurate.

Re: what am I seeing

Posted: Mon May 24, 2010 4:49 am
by wisperer
thank you Lebowski;

When I look at the graph on each client, most will show spike in only one service like http while ping is good or ssh spike while ping is good.

If I am to monitor only one function, I guess it should be ping?

I used solarwinds in the past: nice function was % packet loss for each sub was easily visible

Re: what am I seeing

Posted: Mon May 24, 2010 3:27 pm
by lebowski
There is a packet loss probe in the probe thread...

I typically just let every probe notify me and if I ever see 200 messages I just mark them all read. There is almost no way to monitor devices with out an occasional overflowing inbox. I have stopped trying to reduce the amount of email and just make sure that there are zero false positives.