A server alert arrives while your laptop is out of reach. Before changing anything, answer a smaller question: what can you verify safely from the phone?
We're the team behind TermiGram, an SSH terminal and SFTP client inside Telegram. This is the read-first workflow we'd recommend with any mobile SSH client. The commands below inspect state; they don't restart services, delete files, or change permissions.
Make the first connection before you need it
Use a server you own or are authorized to administer, preferably a non-critical test machine for the first run. Have working SSH credentials and a recovery route outside your phone.
In TermiGram, open @TermiGram_Robot, launch the Web App, choose Standard protection or set up a Master Password, then add a server profile and open Terminal. SSH reachability and authentication still have to work. A Telegram client does not bypass a firewall or repair a broken SSH service.
Use a restricted account appropriate to the task. A harmless-looking command typed into a root shell still has root privileges if you mistype it.
Confirm the machine before trusting the prompt
Run these separately so you can read each result:
hostname id -un Compare the hostname and account with the machine you intended to inspect. A recognizable profile label helps, but it isn't a reason to skip this check. If either result surprises you, stop before continuing.
Keep hostnames, usernames, IP addresses and command output out of public screenshots.
Check space before reaching for a restart
df -h df -i The first command reports space on mounted filesystems in human-readable units. Look at the filesystem used by the affected application, not just the first row.
The second reports inode usage. A filesystem can run out of inodes while it still has free bytes. That is why "there's space left" isn't always enough to rule out a storage problem. Filesystem types differ; an unhelpful inode report isn't proof that storage is healthy.
Don't paste a cleanup command into the next line. A full filesystem tells you where to investigate; it does not tell you which files are safe to remove. Avoid a recursive disk scan as the automatic next step on an already struggling server.
Read uptime and service state as clues
uptime An unexpectedly short uptime can point to a recent reboot. Load averages are not CPU percentages, and a high value alone doesn't identify the cause. Compare with what is normal for that machine.
On a Linux host that uses systemd:
systemctl --failed --no-pager This lists failed units without opening an interactive pager. A failed unit might be an old, unrelated failure. An empty list doesn't prove the website, database or containerized application is healthy.
If the host doesn't use systemd, skip this command. If access is denied, don't change permissions just to make a diagnostic command succeed.
| Observation | What it establishes | What it does not establish |
|---|---|---|
| SSH opens | This account can reach a shell now | The application is healthy |
| A filesystem is nearly full | Capacity needs investigation | Which files are disposable |
| A unit is failed | systemd recorded a unit failure | That restarting it is safe |
| No failed units | No units are listed as failed | That requests or jobs succeed |
Need a file? Take the small, relevant one
Open SFTP from the same server profile and browse to a file your account is allowed to read. Download only what the investigation needs. Don't export an entire application directory because you aren't sure which log matters.
Logs can contain tokens, customer information or connection strings. Check what you're downloading, where the phone will save it, and what must be redacted before sharing.
TermiGram is an SFTP client, not a backup service. Its documentation specifies a 250 MB and two-minute limit for an individual in-app copy operation; keep Telegram open during a transfer. Use a server-side transfer workflow for larger data rather than repeatedly retrying from a weak mobile connection.
Decide whether to continue
Our boundary for phone work is deliberately conservative: collect enough evidence to choose the next action. If a fix needs several commands, a configuration diff, or a rollback plan, moving to a workstation can be the correct result.
A normal terminal tab also isn't a guarantee that a process survives disconnection. Set up and test an appropriate server-side session or service workflow before depending on it.
For a first TermiGram check, keep the scope small: connect to one non-critical server, verify the hostname, read filesystem usage, then disconnect. You don't need to change production to decide whether the mobile workflow suits you.
The continuing Free plan includes one server profile, one Terminal session and one SFTP workspace. New accounts start with a five-day Welcome Trial with higher limits; Cloud Terminal and bulk execution are not part of continuing Free. Current limits are in the plan documentation. Product: TermiGram.
If you try the workflow, which step gets in the way: opening the app, connecting, reading output, or finding a file? Describe the step, not your credentials or server details.
Sources: GNU df manual, systemctl manual, TermiGram Terminal, SFTP, and security boundaries.
AI disclosure: Written by Codex, an AI agent, for the TermiGram project account. Product statements were checked against the current public documentation. This is a suggested runbook, not a report of a customer incident or a benchmark.
