Most problems in Linux do not appear due to a major catastrophe, but rather due to accumulated small things: full disk, hung services, uncontrolled logs, outdated packages.
This post leaves you with a monthly checklist that you can execute in 30-45 minutes.
Category hub for more guides: /blog/category/linux/.
1) Disk and basic cleaning
Start with visibility:
df -h
du -sh ~/* 2>/dev/null | sort -h
Then clean carefully:
- large caches that you don’t use,
- old artifacts,
- excessive local logs,
- deprecated images/containers if you work with Docker.
If you don’t know what to delete, first move it to the temporary folder and validate it for 7 days.
2) Updates without breaking setup
It is not “updating for the sake of updating”. Do it in order:
sudo apt update
sudo apt upgrade -y
Then validate:
- critical tools (git, node, docker),
- audio/network/peripherals,
- local services.
3) Network and connectivity
Quick checklist:
- DNS resolves well?
- stable ethernet/wifi speed?
- Does VPN interfere with local routes?
/blog/18-solution-ethernet-speed-limitation-on-ubuntu-2204//blog/17-setting-default-audio-device-at-session-start-on-ubuntu-2204/
4) Services and processes
Review consumer processes:
ps aux --sort=-%mem | head -10
ps aux --sort=-%cpu | head -10
If there is recurrent abnormal consumption, document the cause and corrective action.
5) Backups and recovery
“I make a backup” is not enough. You should be able to restore.
Checklist:
- dotfile backup,
- backup of critical projects,
- monthly restore test in temporary folder.
6) Basic security
- validate SSH access and active keys,
- check unnecessary open ports,
- keep basic firewall configured.
7) Monthly close (10 minutes)
Leave a brief record:
- maintenance date,
- findings,
- pending actions.
This gives you traceability and reduces reactive debugging.
Closing
Stable Linux doesn’t depend on luck, it depends on routine. A small but constant checklist avoids most major pain.
Recommended next step:
Happy reading! ☕
Comments