Telegram Server Monitor Bot for Linux Servers

What the project does

Telegram Server Monitor connects a Linux server to a Telegram bot through a small local agent. After pairing, the agent sends health data to the backend over HTTPS. The bot can then show the server status, recent heartbeat, CPU usage, RAM usage, disk usage, load average, uptime and detected systemd services.

The important idea is that the operator does not need to give Telegram an SSH password. The backend does not open an inbound SSH session to the server. Instead, the local agent is installed on the server, keeps its own local secret and reports only the data needed for monitoring and allowed maintenance actions.

Who it is for

How the setup works

1. Open the Telegram bot from server.howprog.one.
2. Send /add_server to generate a temporary pairing code and install command.
3. Run the generated command on a Linux server that you own or are allowed to administer.
4. The installer registers the server, creates the systemd service and starts the local agent.
5. The bot sends a confirmation message with server status, detected services and available buttons.

curl -fsSL https://server.howprog.one/agent/install.sh -o install.sh
sudo bash install.sh --api https://server.howprog.one --pair PAIRING_TOKEN

The pairing token in the example is intentionally a placeholder. Real tokens should be short-lived and should not be published in screenshots, docs or public chats.

What it monitors

• CPU and load average: useful for detecting runaway processes, traffic spikes or overloaded PHP, database and mail services.
• RAM usage: useful when a server starts swapping, processes grow unexpectedly or web workloads consume too much memory.
• Disk usage and inodes: important for uploads, logs, caches, mail queues and databases that can stop writing when space is exhausted.
• Uptime and heartbeat: helps confirm that the agent is alive and the server is still reporting.
• systemd service state: shows detected services such as Nginx, Apache, PHP-FPM, MySQL, Redis, cron, mail services or other units visible on the server.

Safe Telegram actions

The project is not only a status bot. It is designed around practical actions that an operator may need during routine maintenance. A server owner can request status, open a services list, restart an allowed detected service, clean allowed logs or reboot the server after confirmation.

These features are powerful, so the safety model matters. The bot should never become a public shell. Actions should be allowlisted, visible, confirmed when risky and limited to predictable maintenance commands. A full server reboot should always require confirmation because it can interrupt websites, mail delivery, databases and background jobs.

Security model in plain language

• No SSH password is needed for the Telegram workflow.
• No private SSH key, hosting-panel password, database password, email password or FTP password should be sent to the bot.
• The agent connects out over HTTPS, so the backend does not need inbound SSH access.
• The pairing code is temporary and is used only to register the server.
• Maintenance commands should be allowlisted instead of accepting arbitrary shell text from Telegram.
• Reboot and destructive actions should require explicit confirmation.

For a deeper security explanation, open the project security page on server.howprog.one/security.

How it differs from classic uptime monitoring

External uptime checks are useful because they confirm whether a public HTTP endpoint responds from outside. Telegram Server Monitor is different: it is agent-based. It can report internal server signals that an external checker cannot see, such as RAM pressure, disk usage, load average, systemd service state and heartbeat freshness.

The best setup can use both approaches. External uptime monitoring catches public reachability problems. The Telegram agent explains what is happening inside the server and gives the operator faster context for the next action.

Useful scenarios

• A website feels slow and you want to check load, RAM and web service state from your phone.
• A disk usage alert fires before logs or uploads fill the server completely.
• A cron or web service stops, and the bot shows a restart button for the detected unit.
• A small team needs shared alert visibility in a Telegram group.
• You manage several small VPS servers and want a quick status command without logging into every machine.

What it should not replace

Telegram Server Monitor should not replace backups, SSH emergency access, provider-level console access, security patching or a full incident response process. It is a fast operational layer: it helps you notice problems, inspect common server health metrics and run routine allowed actions faster.

If a server is compromised, unstable or losing data, the right response is still a proper recovery workflow: isolate the host if needed, preserve logs, rotate secrets, restore from trusted backups and review the root cause.

Project link

The project is available at server.howprog.one. The site includes setup instructions, bot command documentation, security notes, troubleshooting pages and the Telegram bot call to action.

FAQ

Related HowProg.one pages

• HTTP Headers Checker for checking response headers during deployment.
• SSL Checker Command Helper for HTTPS certificate and redirect diagnostics.
• XML Sitemap Indexing Checklist for publishing new website pages cleanly.
• Cron Job Not Running on Linux for debugging scheduled jobs on monitored servers.