How to Install Online Server Monitor on Linux

This guide explains how to add one Linux server to Online Server Monitor, verify the local agent, check logs, configure alert thresholds, reinstall the agent, and remove it when it is no longer needed.

The agent is intended for systemd-based servers used for websites, mail, databases, application backends and small VPS deployments.

• Ubuntu and Debian servers with systemd.
• AlmaLinux, Rocky Linux, CentOS Stream and RHEL-like systems.
• Dedicated servers and VPS hosts with outbound HTTPS access.

You need curl, bash, systemd, outbound HTTPS access to this site, and root or sudo access on the server. The service does not need inbound SSH access and does not ask for SSH passwords.

Open @live_server_monitor_bot in Telegram and send /start. Use a private chat for personal servers or add the bot to an operations group if multiple people should see alerts.

Send /add_server. The bot creates a short-lived pairing token and prints a complete install command. You do not create the token manually; you only copy the command that Telegram shows.

Paste the command into the target server terminal as root or with sudo. Replace PAIRING_TOKEN only with the token generated by the bot. Never publish a live pairing command in a public chat.

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

After the installer finishes, confirm that the agent service exists, is enabled and is running.

systemctl status hp-server-agent
systemctl enable hp-server-agent

If the server does not appear online in Telegram, check recent logs. These commands are read-only diagnostics.

journalctl -u hp-server-agent -n 100 --no-pager
curl -I https://server.howprog.one/healthz

Use the Telegram threshold controls to set CPU, RAM, disk, load and SSL certificate expiry limits. Start with conservative values, then adjust after you see normal server behavior. A threshold should warn you before users notice trouble, without sending noise every few minutes.

The agent looks for public domain names in common nginx and Apache virtual host configuration files, then checks the certificate served on port 443 for each discovered domain. The bot reports the minimum number of days left and can alert before a certificate expires.

Generate a new pairing token with /add_server and run the install command again. Reinstalling is useful after moving a server, rotating the local agent secret, or repairing a broken service file.

Stop and disable the service before removing local configuration. Only run removal steps on a server you control and after you have confirmed the server no longer needs monitoring.

systemctl disable --now hp-server-agent
systemctl status hp-server-agent

• Pairing token expired: generate a fresh command with /add_server.
• curl cannot connect: check DNS, firewall and outbound HTTPS.
• Permission denied: run the install command with sudo or as root.
• Service failed: inspect journalctl -u hp-server-agent -n 100 --no-pager.

Before you paste the install command into a production server, decide which Telegram chat should own alerts, who is allowed to press maintenance buttons, and whether reboot actions should be used during business hours. Monitoring is most useful when the operational rules are clear before the first incident.

• Use a private chat for a personal VPS.
• Use a controlled team group for shared operations.
• Do not install the agent from a forwarded or stale pairing command.
• Keep a second access path to the server, such as SSH or hosting console, for emergency recovery.

The installer downloads or writes the local agent files, creates local configuration, registers the server with the backend, and starts a systemd service. It should not ask for SSH passwords, hosting panel passwords, database credentials, or cloud provider tokens. The pairing token is used only for initial registration.

Use these paths when you need to audit the installation or explain it to another administrator. Exact internals may change as the agent evolves, but the systemd service name is the main operational handle.

• Service name: hp-server-agent.
• Service status: systemctl status hp-server-agent.
• Recent logs: journalctl -u hp-server-agent -n 100 --no-pager.
• Local configuration should stay on the monitored server and should not be posted in Telegram.

Repeat the /add_server flow for every server. Each server should receive its own pairing token and its own local agent secret. Do not reuse one pairing command across several machines because that makes ownership and troubleshooting harder.

• Give servers recognizable hostnames before installing the agent.
• Pair a server from the Telegram chat that should receive its alerts.
• After installation, use /servers to confirm every machine appears once.

The first report is not just a success message. Use it to verify that the bot sees the right hostname, realistic CPU and RAM values, disk usage, load average, uptime, and detected services. If values look wrong, troubleshoot before relying on alerts.

A good rollout is simple: install one server, wait for metrics, tune thresholds, test read-only status commands, then decide whether to expose maintenance buttons. This keeps the first production setup understandable and reduces surprises for the team.

• Read bot commands to understand the Telegram workflow.
• Read security model before enabling service restarts or reboot.
• Keep troubleshooting bookmarked for agent checks.

• The bot does not need SSH passwords.
• The pairing token is temporary and should be copied from Telegram exactly as shown.
• The first server report may take a short time after installation.
• For security details, read the security model.