Applies to: Uptime Kuma 2.5.9 (web interface, English)

In Uptime Kuma, first create a notification (Email (SMTP), Telegram, Slack, Discord or one of the 90+ other services), press Test to check it, then add an HTTP(s) or HTTP(s) - Keyword monitor for your website and tick that notification. Uptime Kuma then sends a message when the site goes down and another when it comes back.

This guide assumes Uptime Kuma is already running. If not, follow installing Uptime Kuma with Docker, use an Uptime Kuma VPS that we install for you, or choose managed Uptime Kuma hosting with no server to look after.

1. Set up a notification

  1. Open Settings, Notifications and press Set Up Notification (the same button is also in the Notifications section of every monitor).
  2. Choose the Notification Type and give it a Friendly Name, for example "Email to ops".
  3. Fill in the fields for that type (below).
  4. Tick Default enabled if new monitors should use it automatically, or Apply on all existing monitors to add it to the monitors you already have.
  5. Press Test. The message "Sent Successfully." means Uptime Kuma delivered it. Then press Save.

Email (SMTP)

  • Hostname and Port of your mail server, for example port 587 or 465.
  • Security: None / STARTTLS (25, 587) or TLS (465).
  • Username and Password of the mailbox that sends the alerts.
  • From Email, for example "Uptime Kuma" <alerts@example.com>, and To Email (several addresses separated by commas). CC, BCC, Custom Subject and Custom Body are optional; leave the last two blank for the default text.

In our lab test the Test button sent an email with the subject "Email alice Testing" (the Friendly Name plus "Testing"). A real outage on a keyword monitor arrived with a subject like this:

[example.com checkout text] [🔴 Down] 200 - OK, but keyword is not in [Example Domain html{color-scheme:light dark;ba...]

If the test fails, check the port and Security pair first (587 with STARTTLS, 465 with TLS), then the mailbox password. Use a mailbox on a different server from the website you monitor, or the alert cannot be sent when that server is down.

Telegram

  • Bot Token: the UI says "You can get a token from https://t.me/BotFather."
  • Chat ID: a direct chat, group or channel ID. The UI's help text says to send a message to your bot and then open https://api.telegram.org/bot<YOUR BOT TOKEN HERE>/getUpdates to read the chat_id.
  • Optional: (Optional) Message Thread ID for a forum topic, Send Silently and Protect Forwarding/Saving.

Slack

  • Webhook URL (required): an incoming webhook from your Slack workspace.
  • Optional: Username, Icon Emoji and Channel Name (to post to a channel other than the webhook's own, for example #other-channel), plus Notify Channel, which alerts all members of the channel.

Discord

  • Discord Webhook URL: the UI says "You can get this by going to Server Settings -> Integrations -> View Webhooks -> New Webhook".
  • Optional: Bot Display Name, Prefix Custom Message (for example a mention) and Suppress Notifications.

We tested email delivery end to end. The Telegram, Slack and Discord fields above are exactly as the 2.5.9 interface labels them, but we did not send live messages to those services; the steps to create the bot or webhook come from the Uptime Kuma interface help text and its notification documentation.

2. Add an HTTP(s) monitor

  1. Press Add New Monitor.
  2. Monitor Type: HTTP(s).
  3. Friendly Name: a name you will recognise in an alert, such as "Shop home page".
  4. URL: the full address, for example https://www.example.com.
  5. Heartbeat Interval: how often to check, 60 seconds by default. Uptime Kuma warns that "Intervals below 20 seconds may result in poor performance."
  6. Retries: 0 by default. With 1 or more, a failed check first shows as Pending and the monitor is only marked Down, with an alert, when the retries also fail. Heartbeat Retry Interval (60 seconds by default) sets the gap between retries. One or two retries stop alerts for a single slow response.
  7. Request Timeout: 48 seconds with the default 60-second interval.
  8. Resend Notification if Down X times consecutively: 0 means one alert per outage; set a number to be reminded while the site stays down.
  9. Under Notifications, tick the notification you created and press Save.

The monitor counts any status in Accepted Status Codes (200-299 by default) as up. In our lab, a monitor for https://example.com showed "200 - OK" with the response time and uptime for 24 hours, 30 days and a year.

3. Add a keyword monitor

A page can return 200 and still be broken: a blank theme, a database error printed as normal text, or a shop with no products. A keyword monitor also checks that a word or phrase is in the page.

  1. Choose HTTP(s) - Keyword as the Monitor Type and enter the URL.
  2. In Keyword, enter text that is always on a healthy page, such as your footer or "Add to cart". The search is case-sensitive.
  3. Tick Invert Keyword to alert when a word appears instead, for example "Error establishing a database connection".

Lab results: with the keyword present the monitor showed "200 - OK, keyword is found". With a keyword that is not on the page, it went Down with "200 - OK, but keyword is not in [...]" and the start of the page. With Retries set to 1, the first failure showed as Pending and the second as Down, and only the Down result sent an email.

Tips

  • Run Uptime Kuma on a different server from your websites, so the monitor stays up when they go down.
  • Add the same site twice if you like: an HTTP(s) monitor for the home page and a keyword monitor for checkout or login.
  • To watch ports, ping, DNS and cron jobs as well, see server monitoring software for Linux compared.

Official documentation: Uptime Kuma notification methods.

Related services

KVM VPS hosting

NVMe SSD and full root access, with Free Basic Managed Support.

Unmanaged dedicated servers

full root access and free IPMI KVM, with Free Basic Managed Support.

Managed dedicated servers

our engineers look after the server for you, from $135 a month.

Was this answer helpful?

Still stuck? Open a ticket