Don't Panic — Network Troubleshooting

Network troubleshooting is the practice of finding where communication fails between a device and a service by comparing working and failing paths one layer at a time. Before anyone formalized this, network problems were solved by restarting things until the symptoms stopped, which is a strategy with an impressive failure rate and an even more impressive ability to hide the actual cause.

The layered model is the load-bearing idea. Communication stacks in order: local interface and link, IP route and gateway, name resolution, transport port, application. You test from near to far, because confirming local state before blaming remote systems saves everyone time. Each layer answers a different question, and a working answer at one layer does not prove the next layer works. ping sends ICMP Echo probes and tells you whether one probe received a reply. It does not prove DNS works, a TCP port is open, or the application answers. traceroute reveals which routers a packet crosses, but a missing hop can be filtering or rate limiting, not a failure.

The part nobody expects is that the network is genuinely somewhere else. You can capture packets on the server and miss the failure because the reply never reached the client. traceroute shows you one direction, not the symmetric path, and asymmetric routing is normal, not exceptional. ICMP can be filtered, so a failed ping with a working TCP connection means a firewall rule, not an outage. DNS resolves the name you typed to an address, and two resolvers can give you different answers at the same moment. Latency and packet loss are not the same thing: one measures how long a round trip takes, the other measures how many probes never came back. A port that reports open tells you the handshake succeeded, not that the application behind it functions.

The discipline that saves time is simple: change one thing at a time, predict what you should see, run the smallest safe test, and compare the result with a baseline. When you escalate, carry a reproducible case with source, destination, protocol, port, timestamp, and the first point where working and failing paths diverge.

Read the Intro for the full diagnostic map and glossary. The Cheatsheet holds the probe interpretation table and scope matrix. The Field Notes cover the operational traps that appear when you think you have narrowed the failure and have not.

Don't Panic — Network Troubleshooting

Relevant careers