Run a Python Bot 24/7 on a DigitalOcean VPS

How to deploy a persistent Python bot on a Linux VPS using a dedicated user, virtual environment, systemd, logs and a firewall.

Keep the server boring

For a small bot, reliability usually improves when the server has fewer moving parts. A common stack is Ubuntu, one dedicated application directory, one Python virtual environment, systemd, and either SQLite or a managed database.

Suggested layout

/opt/bot/
  app/
    main.py
    broker.py
    strategy.py
  data/
    bot.db
  logs/
  venv/
  .env

Install into a virtual environment

python3 -m venv /opt/bot/venv
/opt/bot/venv/bin/pip install -r /opt/bot/app/requirements.txt

A virtual environment avoids mixing bot dependencies with the operating system's Python packages.

Run under systemd

Do not rely on an SSH terminal staying open. systemd can start the process at boot, restart it after a crash, and put stdout/stderr into the journal.

[Unit]
Description=Bot service
After=network-online.target
Wants=network-online.target

[Service]
Type=simple
User=bot
WorkingDirectory=/opt/bot/app
EnvironmentFile=/opt/bot/.env
ExecStart=/opt/bot/venv/bin/python /opt/bot/app/main.py
Restart=on-failure
RestartSec=5

[Install]
WantedBy=multi-user.target

Operations commands

sudo systemctl daemon-reload
sudo systemctl enable --now bot.service
sudo systemctl status bot.service
journalctl -u bot.service -n 100 --no-pager
journalctl -u bot.service -f

Firewall

If the bot does not need a public web server, do not expose one. If it does, put HTTPS/reverse proxy or Cloudflare in front, and bind internal services to localhost or a private interface when possible.