A server alert arrives while you're away from your laptop. Before restarting anything, check three things: the host, disk space, and failed services.
These commands work in any SSH client. For the Telegram workflow, open TermiGram's Mini App, select your server card, and tap Connect. You'll need working SSH access to a server you're authorized to administer. Review the security model before entering credentials.
1. Confirm the host and account
hostname
id -un
Compare both results with the server and account you intended to use. If either is wrong, disconnect before continuing.
2. Check the affected filesystem
df -h
Find the mount used by the affected application. Check its Avail and Use% columns; free space on another mount won't help it.
If an application reports "No space left on device" despite free bytes, check inodes too:
df -i
A filesystem can exhaust its inodes before its storage capacity. A full filesystem is a reason to investigate, not permission to delete unfamiliar files.
3. Look for failed services
On a systemd host:
systemctl --failed --no-pager
If a listed unit belongs to the affected application, inspect it. For example, only if your service is named nginx.service:
systemctl status nginx.service --no-pager
Replace that name with the relevant unit. status shows its current state and recent log lines; it doesn't restart it. An empty failed-unit list doesn't prove the application works, so check the failing request or job as well. Skip these commands on hosts without systemd.
Before making changes, note the affected mount or unit and the exact error. Keep credentials and sensitive log output out of public messages. A restart can wait until you know what you're trying to fix.
References: GNU df, systemctl, Terminal guide.
AI disclosure: Written by Codex for the TermiGram project account.
Top comments (0)