Find out what your application or server is doing.
A log is a record of something that happened in a system.
Applications, servers, databases and services create logs to help you understand:
- What happened
- When it happened
- Whether something went wrong
- What caused a problem
Think of logs as the history of your system.
1. What Does a Log Look Like?
A simple log might look like:
2026-10-04 01:20:15 INFO User 42 logged in
2026-10-04 01:21:03 WARN Database response is slow
2026-10-04 01:21:10 ERROR Database connection failed
Look for:
TIME
LEVEL
MESSAGE
2. Log Levels
You'll commonly see:
INFO
Something normal happened.
INFO User logged in
Usually nothing to worry about.
WARN
Something unusual happened.
WARN Disk space is getting low
Pay attention, but it isn't necessarily a failure.
ERROR
Something failed.
ERROR Database connection failed
This usually needs investigation.
DEBUG
Detailed information useful when troubleshooting.
DEBUG Sending request to payment service
Usually more detailed than you need during normal operation.
3. Read the Time First
When investigating a problem, find out:
When did it happen?
Look at the timestamp.
For example:
01:20
01:21
01:21
01:22
Then compare the log with when the user reported the problem.
4. Look for ERROR
A quick first step:
grep "ERROR" application.log
You can also search for:
grep "WARN" application.log
Or:
grep -i "error" application.log
-i makes the search case-insensitive.
5. Read the Lines Around the Error
The error itself may not explain the real problem.
For example:
INFO Starting database connection
WARN Database response slow
ERROR Database connection failed
ERROR Unable to load customer records
The first error may be more useful than the last one.
Don't immediately jump to the final error message.
6. Follow a Log as It Happens
For a log file that is actively being written:
tail -f application.log
Now new log entries will appear as they are created.
Press:
Ctrl + C
to stop.
This is extremely useful when:
- Restarting an application
- Testing a feature
- Reproducing an error
- Watching a deployment
7. See the Last Few Lines
tail application.log
More lines:
tail -n 50 application.log
This means:
Show the last 50 lines.
8. See the Beginning of a Log
head application.log
Or:
head -n 50 application.log
9. Search for Something Specific
Use grep:
grep "login" application.log
Examples:
grep "payment" application.log
grep "database" application.log
grep "timeout" application.log
grep "failed" application.log
10. Combine Commands
You can pipe commands together.
For example:
tail -n 100 application.log | grep "ERROR"
Meaning:
Get the last 100 lines, then show only the errors.
Another useful example:
grep -i "error" application.log | tail
11. Logs May Be Somewhere Else
Common places on Linux include:
/var/log/
You might see:
/var/log/syslog
/var/log/auth.log
Applications may also have their own directories:
/var/log/myapp/
Or logs may be managed by the system rather than stored in a normal file.
12. journalctl
Many Linux services use systemd.
To see service logs:
journalctl -u myapp
Follow them live:
journalctl -u myapp -f
See recent logs:
journalctl -u myapp --since "1 hour ago"
13. The Most Important Question
When something breaks, don't just search for:
ERROR
Ask:
What happened?
ERROR Database connection failed
When?
01:21:10
What happened immediately before it?
WARN Database response is slow
What happened immediately after?
ERROR Unable to load customer records
Now you're starting to trace the problem.
14. Common Things to Search For
When troubleshooting, try:
error
failed
exception
timeout
denied
refused
connection
database
authentication
permission
Don't assume the word ERROR will always be present.
15. A Simple Troubleshooting Workflow
When someone says:
"The application isn't working."
Try:
1. Reproduce the problem
↓
2. Note the time
↓
3. Open the logs
↓
4. Look around that time
↓
5. Find ERROR / WARN
↓
6. Read the surrounding lines
↓
7. Identify the first failure
↓
8. Investigate that failure
16. Don't Panic at Every Error
A log can contain errors that aren't the actual cause of the problem.
For example:
ERROR Request failed
might simply be a consequence of:
ERROR Database unavailable
The database error may be the real problem.
Look for the first meaningful failure.
Essential Commands
| What you want | Command |
|---|---|
| Show log | cat application.log |
| Last lines | tail application.log |
| Last 50 lines | tail -n 50 application.log |
| Follow live | tail -f application.log |
| First lines | head application.log |
| Search | grep "text" application.log |
| Search ignoring case | grep -i "error" application.log |
| Service logs | journalctl -u service |
| Follow service logs | journalctl -u service -f |
The Golden Rule
Don't just look at the error. Look at the story around the error.
When reading logs, think:
WHEN?
↓
WHAT HAPPENED?
↓
WHAT HAPPENED BEFORE IT?
↓
WHAT FAILED FIRST?
↓
WHY?
You don't need to understand every line in a log.
Find the time, find the problem, follow the trail.
