How to troubleshoot anything

Nobody in IT knows every fix. What separates good techs from great ones is a method: a way to narrow down any problem, even one you’ve never seen. Learn this, and every other lesson on this path gets more useful.

1. Understand the actual problem

“The internet is down” can mean a dozen different things. Before touching anything, ask:

• What exactly happens? What’s the exact error message? (Get a screenshot.)
• What should happen instead?
• When did it start? What changed around then: an update, a new device, a move, a power outage?
• Who is affected? One person, one room, or everyone?
• Can you make it happen again?

“Everyone, starting this morning” points somewhere completely different from “just me, only on this one website.”

2. Check the simple things first

Experienced techs aren’t too proud for the basics, because the basics are usually the answer:

• Is it plugged in, turned on and connected?
• Has it been restarted?
• Is there a pending update, or did one just install?
• Is the account locked out, or did a password just expire?

Restarting fixes a surprising number of problems. But if it keeps coming back, a restart only hides it. As the saying goes, problems that go away for free come back for free.

3. Divide and conquer

This is the heart of the method. Split the problem in half, test which half is broken, and repeat.

Say one website won’t load:

1. Do other websites load? If not, it’s the connection, not the site.
2. Does it load on another device on the same network? If so, it’s that one computer.
3. Does it load in another browser? If so, it’s the browser, not the computer.
4. Does ping work to an address but not a name? Then it’s DNS (see step 3).

Each question cuts the possibilities roughly in half. In four or five questions, you’ve narrowed a mystery down to one thing.

4. Change one thing at a time

If you change three settings at once and it starts working, you don’t know which one fixed it, or whether the other two broke something else. Change one thing, test, then decide what to do next. Write down what you change, so you can undo it.

5. Read the logs and error messages

Computers usually tell you what’s wrong; people just don’t look. On Windows, Event Viewer (search for it in the Start menu) records errors and warnings with timestamps. Search the exact error text or code online. Someone else has almost always had the same problem.

6. Find the root cause

Once it works again, ask why it broke. A full disk, an expired certificate, or a setting changed during an update will strike again unless you fix the cause. This is the difference between a quick fix and a real one.

7. Document it

Write down the symptom, the cause and the fix, even in a simple notes file. The next time it happens, you’ll solve it in minutes instead of hours. This habit is a huge part of what makes experienced admins look like they know everything.

Know when to escalate

Part of the method is knowing your limits. If something involves possible data loss, a security incident, or a system you don’t understand, stop and get help before making it worse. Asking early is a professional skill, not a weakness.

Try this

Next time something breaks, at home or at work, write down the answers to the questions in the first section before you touch anything. Then try divide and conquer. You’ll be surprised how often the answer appears by the third question.

Key takeaways

• Understand the problem first: what, when, who, and what changed.
• Divide and conquer: each test should cut the possibilities in half.
• Change one thing at a time, and read the error messages.
• Fix the root cause and write it down.