Yet another uptime check tool: u(ptime) + nagi(os) + kuma(Uptime Kuma).
Best of both worlds: one file on one machine works out of the box, and when you want more, join other machines and it becomes a cluster. Same program, nothing to migrate.
Friendly, self-hosted monitoring: know the moment your websites and servers go down, without the noise.
English · 中文 · हिन्दी · Español · العربية · Français · বাংলা · Português · Русский · Bahasa Indonesia · اردو · Deutsch · 日本語 · Tiếng Việt · 한국어
- Ready in about 30 seconds. Download one file, run it, open your browser.
- One file, nothing else to install. Works on Linux, macOS, Windows and FreeBSD.
- Checks what matters. Websites, servers, ports, ping, DNS, SSL certificates and domain expiry.
- Alerts where you already are. Email, SMS and calls; Feishu, DingTalk, WeCom, Telegram, Slack, Discord, Teams, LINE, Mattermost, Google Chat; PagerDuty, Opsgenie, Datadog; ntfy, Gotify, Bark, Pushover, Pushbullet, ServerChan, Apprise; and any webhook. Each can have its own message template.
- Status pages to keep your users in the loop.
- Easy to switch. Bring your monitors over from Uptime Kuma, Nagios, Checkmate or Gatus, in the web page or on the command line.
- Seven themes, or your own. Five tones from white to black (Paper, Mist, Slate, Graphite, Ink) plus Sky and Violet, with Nagios-style status colours; upload a theme pack for your own colours, and with a registration your own product name and logo.
- Speaks your language. The interface comes in 15 languages.
1. Download the file for your system from Releases and unpack it.
2. Run it:
./unagikumaOn Windows, run unagikuma.exe.
3. Open http://localhost:8077 in your browser and create your admin account.
4. Click Add monitor, enter a website address, and you're done.
Opening it from another computer? Use
http://<server-address>:8077and enter the Setup token that the program printed when it first started.
- One machine: run the file, add a monitor, done. No database, no agents.
- Two machines: paste a join code under Run mode. Each machine checks and alerts on its own, so either one can go down without silencing your alerts.
- A cluster: add as many sites as you like. Every alert shows what each site sees, so you can tell "down everywhere" from "down from here".
- Your own network is down? One message about the network, not one per monitor.
- A host is down? Its services are not reported one by one.
- At most three alerts a minute per channel; the rest arrive as one summary. Recoveries always go out at once.
In the web page, open Configuration › Import, pick Uptime Kuma, Nagios, Checkmate or Gatus, upload the file, check the preview and add it.
Or on the command line: look first with --dry-run (nothing is changed), then run the same command without it.
./unagikuma import kuma --dry-run /path/to/kuma.db
./unagikuma import nagios --dry-run /etc/nagios4/nagios.cfg
./unagikuma import checkmate --dry-run monitors.json
./unagikuma import gatus --dry-run config.yamlCheckmate and Gatus: HTTP checks are moved. Anything that cannot be moved exactly is listed and stops the import, so nothing is half-done.
On the machine where Unagikuma runs:
./unagikuma reset-password adminType the new password twice (it is not shown). Add --generate to get a random one instead.
- Contacts. E-mail and SMS go only to active people in the directory. The channel Test button uses that same list by default, so a test cannot look successful when nobody would receive a real alert.
- Retries. A failed notification channel is retried a couple of times before fallback (override with
notify.channel-retries=0to keep the old no-retry behaviour). Probe retries are separate. - Listen. The UI defaults to plain HTTP on the listen address (often
:8077). Prefer a TLS reverse proxy or--tls-cert/--tls-keyfor anything beyond a private network. Agents already use pinned TLS on the same port.
- User manual — everything else, step by step
- Changelog
Apache License 2.0, see LICENSE.