I have an extensive home automation setup and multiple computer users at home - 100 nodes roughly.

Once every month or so the network grinds to a halt and I can't put my finger on it. It happened last night when the my switch (Unifi Dream Machine Pro) reported that it could no longer connect to the Internet (FIOS direct ethernet). I checked FIOS and it was running fine. I then unplugged both switches and the UDMP was happy. Then I started plugging the switches and the devices back in one by one until I find what I think is the cause. But invariably it will be something different each time and often I have issues for a few hours and the connection I thought was the issue might not be the one.
So in my limited network knowledge this tells me that some device is potentially jamming up the network but I can't seem to pinpoint it to a single device, in fact it may be more than one.
I downloaded PingPlotter as several sites said it can be used to find out what is slowing down your network.
Ideally I can get something that says this IP is doing something bad and warns me so I can track it and take any corrective measures before everything stops responding. Thanks.

Hey Simon,

Thanks for checking out PingPlotter.

PingPlotter at its core is the traceroute utility on steroids. It should help you in your troubleshooting once you get it set up and monitor all the devices on your network (which seems pretty extensive with 100 different nodes). It would also help to have it monitor traffic leaving your network as well. PingPlotter also does have an alert feature that you can set up for a variety of circumstances.

I would start by using the Local Network Discovery tool in PingPlotter to find every device connected to your network (Tools --> Scan for Local Devices). Then start a trace to all of these devices by hitting the Add All to Summary button in the bottom right. This will start a PingPlotter trace to every device. You’ll also want to start a trace to any public service, like 8.8.8.8 Google DNS.

The next step is to wait for your network to go down. This will likely show up in PingPlotter as a bunch of packet loss (red) and/or high latency spikes (black). Now for the “fun” part. Much like you did previously, you’ll need to start disconnecting your devices from the network one at a time and see if the network improves. You’ll know the network is improving in PingPlotter when you see the red start to disappear. If things aren’t improving, reconnect the device, and move on to the next. Do this until you find the one that is causing the network to fail.

I know this is a bit of a process but there isn’t really a magic silver bullet. Here are some articles you can read through to give you more details and ideas on how to proceed:

A Broadcast Storm Defined and How to Fix It
Find The Device Slowing Down Your Network
A Broadcast Storm Defined & How to Fix It

If you get some data and want to pick our brains on your situation you can reach out to us here on the forums or at support@pingman.com . If you have any specific questions about PingPlotter let me know and I’ll answer any questions you have.

Thanks - I am following the advice below and starting to analyze.


Will report back.

Following up on this - it has happened twice in the last several days. Both times I ran pingplotter and went through the process of unplugging cables. Both times it was inconclusive. I thought I had the suspect, unplugged it and everything returned to normal, plugged it in and things went haywire, but then everything cleared up and plugging and unplugging these connections made no difference.

I did get Wireshark running during the second issue.

I did see a lot of these:

[Reassembly error, protocol TCP: New fragment overlaps old data (retransmission?)]

This was between the machine running Wireshark and my main desktop - a Mac Studio. I was running Anydesk on the Mac to connect to the windows PC running Wireshark. So not surprised that things were going wrong between the two when the network was getting overloaded. Presumably this is a symptom of the problem - not the cause?

Most active IPs at the time behind these two machines in order:

Philips Hue hub - unplugged and didn't solve it
A VM running on one of the two machines above. I didn't try shutting that down but I don't think that is the issue.

Is there anything more I can do on pingplotter to help diagnose? Right now I am in wait and see mode until it happens again and hope I can grab more data. Thanks.

Hey Simon,

The way you're using PingPlotter is the best method for troubleshooting your issue.

It appeared from your post as though you had your suspicious devices figured out.

"I thought I had the suspect, unplugged it and everything returned to normal, plugged it in and things went haywire, but then everything cleared up and plugging and unplugging these connections made no difference."

It seems as though this device was an issue till you unplugged it, and returned to cause an issue when you plugged it back in. Everything clearing up afterwords is great but that device still caused those issues.

I recommend seeing if that device has firmware updates or if you can leave it off your network and see if it overall improves.

If you get some more data you want us to take a look at feel free to post it here.

Thank you,