packet loss in timeline graph are 100% spikes or 0%

My plots have 100% spikes everywhere, but the numbers in the data area show the averages corectly (like 7% etc). I want to see the moving averages like in the graphs showing rising and falling red areas.

Maybe I'm just reading it wrong too. <img src="/forums/images/graemlins/blush.gif" alt="" />

I've included plots for any other thoughtful comments. I'm especially interested in the bottom three plots. What am I seeing here? The last hop is wireless over a Fixed Wireless DSL connection with an access point that is overloaded by virus ridden computers during peak times, but I wonder about hop 14 too.

Thanks!

Changing the packet loss range on the plots from 100% to 5% doesn't change anything. There are still just a bunch of spikes.

Thanks.

Expanding the third from the bottom looks really interesting. What is happening here?

Ok. Just figured out that adjusting the scale to a large value (24 hours, etc.) that the spikes are averaged out. Sorry!

Still interested in your input on that wacky plot though! <img src="/forums/images/graemlins/laugh.gif" alt="" />

Thanks,
Jerrill

You figured out the 100% packet loss part of things there.

See here for some information on that:

http://www.nessoft.com/kb/20

Also, the latency variants you're seeing at hop 14 are not terribly abnormal. We discuss that topic here:

http://www.nessoft.com/kb/5

and

http://www.nessoft.com/kb/24

Of course, none of this directly addresses that packet loss you're seeing across the board there. It's difficult to tell precisely where the packet loss is being initiated - but it's somewhere around hops 2 or 3. Hop 2 shows some packet loss, Hop 3 shows more. If both of these IPs are owned by the same ISP (probably your primary ISP), it probably makes sense to contact them and ask about this packet loss problem. It's pretty clear that your problems are not being caused by the downstream hops here - since hop 3 and hop 16 have very similar packet loss counts.

I hope this helps.

- Pete