Work from the bottom of the stack upward and stop at the first layer that fails. Each rung has a quick test and a decision: if it passes, move up. If it fails, fix that layer before testing the next. The commands are for Windows (macOS and Linux equivalents are in the table at the end).
The ladder
| # | Layer | Question | Test | Healthy result |
|---|---|---|---|---|
| 1 | Physical | Is there a link? | Cable seated, link lights, Wi-Fi icon, try another port or cable | Link light on, adapter shows Connected |
| 2 | Adapter | Is the adapter up with an address? | ipconfig /all |
IPv4 address in the expected range, subnet mask, default gateway, DNS servers |
| 3 | Local IP | Is the stack working? | ping 127.0.0.1, then ping own IP |
Replies |
| 4 | Gateway | Can I reach the router? | ping <default gateway> |
Replies, low latency |
| 5 | Remote IP | Can I reach the internet by IP? | ping 1.1.1.1 or ping 8.8.8.8 |
Replies |
| 6 | DNS | Can names be resolved? | nslookup example.com |
Returns an address |
| 7 | Transport | Is the port open? | Test-NetConnection example.com -Port 443 |
TcpTestSucceeded : True |
| 8 | Application | Does the app work? | Browser, app logs, curl -I https://example.com |
Expected HTTP status |
Reading the results
Step 2: wrong address
169.254.x.x(APIPA): the device did not get a DHCP lease. Check the cable or Wi-Fi, VLAN, DHCP scope exhaustion or the DHCP server. Tryipconfig /releasethenipconfig /renew.- Address outside the expected range or a different subnet: a rogue DHCP server, wrong VLAN or a static address left on the adapter.
- No default gateway: gateway is not offered or the adapter is set manually.
- Duplicate address warnings point to a static IP conflict.
Step 4: gateway unreachable Wrong VLAN or SSID, bad cable or port, switch port disabled, Wi-Fi signal or captive portal. Some routers block ICMP, so also test by opening the router page.
Step 5 passes but step 6 fails
The connection works but DNS does not. Compare nslookup example.com against nslookup example.com 1.1.1.1. If the second works, the configured DNS server is the problem. Check DNS servers on the adapter, DHCP options, and VPN DNS settings. Flush the cache with ipconfig /flushdns.
Step 6 passes, step 7 fails The name resolves but the port is not reachable: firewall, proxy, the service is down, or the wrong address (stale DNS). Test from another device to see whether it is only this client.
Intermittent drops
Run a continuous ping (ping -t <target>), look for loss or high latency, and test with a wired connection to separate Wi-Fi issues. Check channel, signal, roaming, duplex mismatch, bad cable, and VPN reconnects. Use pathping <target> for per-hop loss.
Name resolution specifics
- Resolve an internal name:
nslookup server01 <internal-dns>. A "Non-existent domain" answer means the record is missing. A timeout means the server is unreachable. - Check the DNS suffix search list when short names fail but FQDNs work.
- A stale hosts file entry can override DNS:
C:\Windows\System32\drivers\etc\hosts. - Use
Resolve-DnsName example.com -Type A -Server 1.1.1.1for richer output.
Wi-Fi checklist
- Confirm the correct SSID, that Airplane mode is off, and the radio switch is on.
netsh wlan show interfacesshows SSID, signal, radio type, channel and authentication.- Forget and re-add the network if the credentials changed (
netsh wlan delete profile name="<ssid>"). - Update the driver, and check the power-saving setting on the adapter.
- For enterprise Wi-Fi (802.1X), check certificates, the password and device compliance, and whether the user account is locked.
VPN checklist
- Confirm basic internet access works before VPN.
- Check the client version, credentials and MFA prompt.
- After connecting, check the assigned address, routes (
route print) and DNS servers. - Confirm split-tunnel behavior and which resources are expected over VPN.
- Common failures: blocked by a hotel or public Wi-Fi, expired password, time skew, outdated client, ISP or firewall blocking the VPN port.
Evidence for escalation
Include: output of ipconfig /all, the failing step, tracert <target> or pathping, times and scope (which users, which locations), whether the issue exists on wired and wireless, and what changed recently.
Command equivalents
| Task | Windows | macOS | Linux |
|---|---|---|---|
| IP configuration | ipconfig /all |
ifconfig, networksetup -getinfo Wi-Fi |
ip addr, nmcli |
| Renew DHCP | ipconfig /release and /renew |
System Settings > Network > Renew DHCP Lease | sudo dhclient -r, sudo dhclient |
| DNS lookup | nslookup, Resolve-DnsName |
dig, nslookup |
dig, resolvectl query |
| Flush DNS | ipconfig /flushdns |
sudo dscacheutil -flushcache; sudo killall -HUP mDNSResponder |
resolvectl flush-caches |
| Trace route | tracert |
traceroute |
traceroute, mtr |
| Port test | Test-NetConnection -Port |
nc -vz host port |
nc -vz host port |
| ARP table | arp -a |
arp -a |
ip neigh |
| Routes | route print |
netstat -rn |
ip route |
Safety
Only test systems you are authorized to troubleshoot. Do not run port scans on networks you do not manage without permission.