When a network starts feeling slow, it is easy to assume that the switch itself needs to be replaced. In my experience, though, it is worth checking whether the problem is actually related to traffic congestion before buying new hardware.
One thing I usually look at is how much traffic is passing through the switch during the periods when users notice performance problems. A switch may have plenty of available ports, but that does not necessarily mean it has enough capacity for the traffic moving through its uplinks.
I also think it is important to look at the connections between access switches and the core network. If several busy devices are sharing a relatively limited uplink, congestion can occur even when the individual access ports appear to be working normally.
Another area I check is the type of traffic involved. Cameras, access points, servers, backups, and file transfers can all create very different traffic patterns. A network that works fine during normal office use may behave differently when several high-bandwidth devices are active at the same time.
Monitoring can help identify whether the issue is continuous or only happens during specific periods. Looking at interface utilization, errors, dropped packets, and uplink activity can provide more useful information than simply replacing the switch and hoping the problem disappears.
I also try to consider the network as a whole. The switch, uplink, transceivers, cabling, router, and connected devices all need to support the expected traffic levels. Sometimes the limitation is somewhere other than the switch itself.
For anyone who has dealt with network congestion, what measurements do you normally check first?
Do you start with switch interface utilization, uplink traffic, packet drops, or the connected devices?
I've also been comparing different networking equipment and infrastructure options through WAVOnline while looking at possible upgrade paths.