Site owners are often the last to learn their site is down: a customer calls, a manager messages, and only then does anyone look for the cause. Orders, calls and trust are lost in the meantime. Monitoring flips that order by alerting you the minute the site stops responding. Uptime Kuma is a free, open-source, self-hosted monitoring tool with a simple interface that takes minutes to set up.
What Uptime Kuma can watch
Each check is called a monitor. HTTP(s) monitors fetch a page on a schedule and record the status code and response time. Keyword monitors check that a specific word appears, which catches pages that return 200 but display "Database Error". You can also watch TCP ports such as a mail server's 25 or 993, ping, DNS records and Docker containers. HTTPS monitors can warn you before an SSL certificate expires. Push monitors work the other way around: your script calls Uptime Kuma regularly, which is ideal for cron jobs and backup scripts.
Where to install it
Never run monitoring on the same server as the sites it watches, because if that server dies, so do the alerts. A small separate VPS, a server elsewhere or an always-on office machine works well, and Uptime Kuma needs very few resources. Even the smallest VPS handles dozens of monitors. Data lives in a separate volume, so updating the container keeps your settings and history, and in the example below the port is bound to localhost because outside access goes through a reverse proxy with HTTPS.
Installing with Docker
docker volume create uptime-kuma
docker run -d --name uptime-kuma --restart=always -p 127.0.0.1:3001:3001 -v uptime-kuma:/app/data louislam/uptime-kuma:1
# /etc/caddy/Caddyfile
status.example.uz {
reverse_proxy 127.0.0.1:3001
}
The project is actively developed and new major versions may appear, so check the image tag in the official docs. Create the admin account on first visit and enable two-factor authentication, since the dashboard knows the addresses of all your services.
Telegram alerts
Uptime Kuma supports email (SMTP), Telegram, Slack, Discord, webhooks and more. In Uzbekistan Telegram is the most practical. Create a bot via @BotFather, send it a message and find your chat ID:
curl "https://api.telegram.org/bot<TOKEN>/getUpdates"
# the "chat":{"id": ...} value is your chat ID
Under Settings, Notifications, choose Telegram, enter the token and chat ID and press Test. A group chat lets the whole team see alerts, and marking the notification as Default enabled attaches it to every new monitor.
Example: a clinic site with online booking
A dental clinic in Tashkent takes bookings on misol.uz. Its admin set up an HTTPS check of the home page every 60 seconds, a keyword monitor on the booking page, an HTTP monitor for the API and a TCP monitor for the mail server's port 993. The nightly backup script calls a push URL when it finishes:
curl -fsS -m 10 "https://status.example.uz/api/push/AbCdEf123?status=up&msg=OK" > /dev/null
With a 25-hour push interval, a missed backup triggers a Telegram alert by morning. After one plugin update the site began returning a blank page with code 200; a plain HTTP check would have missed it, but the keyword monitor alerted within two minutes and the issue was fixed before patients noticed.
Status pages, maintenance and false alarms
A public status page shows selected services and recent uptime, so customers can see a known issue instead of calling. Maintenance mode pauses alerts during planned work. To avoid alert fatigue, set Retries to 2-3 so a monitor goes DOWN only after several consecutive failures, use 60-second intervals for key pages and 5 minutes for secondary services, and watch response-time graphs, since slowdowns often precede outages. Whether your site runs on sayt.uz hosting or a VPS, external monitoring tells you first and gives support exact times and error types.