Intermittent Ethernet drops in new structured wiring: where should I test first?

I’m troubleshooting intermittent link drops on a new-home structured wiring system and want to isolate the fault before replacing switches or asking the installer to rerun cable.

The affected runs are Cat6, approximately 60 to 90 feet, from a central patch panel to wall jacks. The problem appears on one endpoint at a time, but it isn’t yet clear whether the failure follows the room jack, the patch-panel port, or the connected device. In some cases the Ethernet link light changes state when the connection drops; in others, the VLAN becomes unreachable while the physical link appears to remain up. That distinction seems important for separating a physical-layer fault from switch negotiation or configuration trouble.

The system includes a managed switch and at least one PoE-powered endpoint. I’m also trying to determine whether the drops correlate with PoE load, although I don’t want to assume that’s the cause just because a camera or access point is involved.

For those who troubleshoot structured cabling professionally, would this be the most useful sequence?

- Test the suspect run with a certified cable tester, including wire map, length, insertion loss, return loss, and any intermittent fault indication.
- Substitute short, known-good patch cables at both the patch panel and endpoint, then move the same device to a known-good switch port.
- Move the suspect horizontal cable to another patch-panel and switch port to see whether the problem follows the cable, port, VLAN, or endpoint.
- Check switch logs and port counters for link renegotiation, CRC errors, symbol errors, PoE events, or excessive power draw.
- Repeat the test with PoE disabled or with a non-PoE device, if practical.

How much weight would you give a basic wire-map result if the link still drops under load? Could a marginal termination or damaged cable pass continuity but fail at gigabit speeds, while a defective patch-panel jack produces similar symptoms? Also, if the link lights never change but only one VLAN stops passing traffic, would you stop treating this as a cabling fault and focus on the switch configuration or uplink?

I’m especially interested in what test evidence you’d want documented before concluding that the patch-panel hardware, switch negotiation, or the in-wall Cat6 is at fault.