How to diagnose problems in a local network with several levels of intermediate switches (hubs)?

Question

Windows

The local network is represented by a server with OS Windows Server 2008 R2 Standard, which is connected to the main hub (D-Link DES-1026G). Then there are other computers with OS Windows XP and older that communicate with the server by connecting directly to the hub described above by a twisted pair cable, or through one more intermediate hubs (there are three floors in the building, there are hubs for the floors, including, from them the wire already goes to hub next to the server). The problem is that in this local network, computers are periodically cut off from the Internet, either one or a whole bunch at once. Suspicion is the main thing on the hubs, it can overheat. The problem is that it's not as easy to find as I think. In any case, right there, on the Toaster, I am looking for a program for pinging the IP addresses of machines in order to detect in the event of a jamb of the hub that all the machines behind it were killed. This is the first option. The second option is to send garbage bags to the hub from some infected computer. I may be talking nonsense, but perhaps this is the viral activity of some of the malware hiding on one of the computers, and which "bends" the hub closest to it along the route to the server.
In any case, I'm interested in the experience of colleagues who also administer a large number of machines on the local network, how do you diagnose and fix problems with the availability of machines? What software/hardware do you use?

Answer 1

also administers a large number of machines in the local network

How many hosts are on your network?
I will assume (since it is built on unmanaged switches and a web developer is trying to administer it), your network is actually small. If so, use the free program Friendly Pinger, in which the pinger is friendly, visual, with alerts. What you need.

Answer 2

A network of 300+ ports, a complex of buildings on an area of ​​2 square kilometers
Cisco 2960-S + by traffic aggregation points in the building.
Cisco 4500 in server room
Connection to aggregation points via optics 2x10 gigabit.
SNMP and syslog configured. Monitoring via nagios.
All hardware is managed , that's the main advantage. Intermediate hubs-soap trays are thrown out as a class.
The main problem of freezing soapbox hubs is poor power supply. Set up uninterruptibles.
I note that the power supplies for soap dishes will need to be replaced from linear to pulsed ones.
We started 12 years ago with an 8 port intel hub (which is still running in one of the redundant copper loops)
The fight against hanging cords lasted 8 years in a variety of ways (power supply through an uninterruptible power supply and replacing the power supply with an ATX is best).
We kicked out the last hub from the client network 4 years ago. Since then, the "everything is gone" problem has disappeared for users.

Answer 3

Well, firstly, a hub and a switch are radically different devices. If your network is built on hubs, I feel sorry for you, it cannot be saved.
But something tells me ...
Anyway...
In the first place, most of the problems in your network are hardware related. Your switches are not managed, without support for STP protocols and without loop detection, i.e. an orgy will be created in the network, and the equipment will silently send this orgy and storms to their ports.
Choose L2 \ L2+ managed equipment.
The second problem is with the small switches that users have. You need to get rid of them first of all, bringing all the links from the end equipment to the managed switches as much as possible. A small switch is installed only as a last resort, and, preferably, also managed (there are also with a small number of ports, but worthy).
Well, then, it’s not very clear who your router is, how addresses are distributed, who DNS is, and much more.