Skip to content
Hyperion

Set up alert channels

Hyperion raises alerts in three places: the notification bell in the panel, email, and Slack. The bell needs no setup. Email needs an outgoing mail relay and recipient addresses. Slack needs an incoming webhook. You set the shared channels on Settings, Alerts, and the per-site ones on each hosting’s Monitoring & logs tab.

Before you start

Section titled “Before you start”
  • You are an Owner to save. An Administrator can open the page and sees the forms disabled.
  • For email, the node’s relay is on. A node with its relay off sends no email, and the Recipients row for it shows a relay off pill.
  • For Slack, you have an incoming webhook URL from Slack, Apps, Incoming Webhooks.

Steps

Section titled “Steps”

Connect Slack

Section titled “Connect Slack”
  1. Open Settings in the sidebar, then Alerts under Communication.
  2. In the Slack group, paste the URL into Incoming webhook URL and click Connect. The pill reads connected.
  3. Click Send test to post “Hyperion test message — your Slack webhook is working.” to the channel.

Once a webhook is stored, the page never shows it again, and there is no URL field. To change it, click Disconnect, confirm, and connect the new one. Disconnecting stops Slack posts at once.

On a cluster, only the master posts to this webhook. A worker’s alert reaches Slack only through a site’s own webhook.

The default webhook receives the Slack messages for disk-quota events, billing and care-plan due reminders, new-hosting announcements and WordPress sign-up flood warnings. A hosting’s profile can carry its own webhook, which replaces the default for billing, care-plan and new-hosting messages. The monitor’s down and up alerts use the site’s own webhook only, described below.

Choose who gets email

Section titled “Choose who gets email”
  1. In the Email group, find the Recipients card. It lists one row per node, the master first. Each node mails its own alerts to its own list.
  2. In Addresses, enter the addresses separated by commas, semicolons or new lines.
  3. Click Save on that row.

A row shows no addresses, 1 address or N addresses, and unknown when the node did not answer. A node that did not answer gets no form, so a blank field cannot erase its list.

Hyperion sends an alert that is not addressed to a particular person to every address in the list, and also to the alert addresses of the site’s profile. A letter addressed to a customer goes to that customer only. Entries without an @ are skipped.

Add alert channels to one site

Section titled “Add alert channels to one site”
  1. Open Hostings, click the site, then open its Monitoring & logs tab.
  2. Under Alert channels, fill Email recipients, Slack incoming webhook URL or Generic JSON webhook URL. See Monitor a site’s availability.

Choose the alert language

Section titled “Choose the alert language”
  1. In the Language & wording group, find Language.
  2. Choose English or Čeština under Alerts are written in. The default is English.
  3. Click Save.

This is the language of the notification bell and of alert email to administrators. Customer letters have their own language setting, on Settings, Customer letters.

Reword the Slack and email text

Section titled “Reword the Slack and email text”
  1. Under Wording, edit any of the three fields. The preview under each fills in sample data as you type.
    • Slack message accepts {message}, {time} and {panel}.
    • Email subject accepts {subject} and {time}.
    • Email body accepts {body}, {time}, {panel} and {kind}, where {kind} is the alert type such as backup.
  2. Click Save.

An empty field sends the alert unchanged. A placeholder Hyperion does not know is sent as typed, so a typo shows in the preview. The wording wraps every Slack message sent through the default or a profile webhook, and every operator email. It does not wrap the two customer letters (the care report and the expiry warning), and a site’s own Slack monitor webhook posts the alert text as is. The sentences of the alerts themselves are reworded with the customer letters, on Settings, Customer letters.

Saving any of these settings restarts the agent of the node being saved in about 5 seconds, and the page says so.

Check that it worked

Section titled “Check that it worked”
  • Slack shows the pill connected, and Send test posts to the channel.
  • Each node’s row under Recipients shows the addresses and no relay off pill.
  • Send a test message from Settings, Mail, using one of the addresses. The message appears in the Email log with state Sent.
  • The wording preview matches what arrives.

If it fails

Section titled “If it fails”
  • “No Slack webhook is configured — save one first.”: connect a webhook, then test.
  • “Slack refused the test message: …”: Slack rejected the post. Check that the webhook is still active in Slack, then disconnect and connect it again.
  • “The agent’s settings could not be read (…). These are defaults, not what is set, so saving is off until it answers.”: the master’s agent did not answer. Reload after it recovers.
  • “Its relay is off, so it sends no email and nobody listed here hears from it.”: turn the relay on from the link on that row.
  • A monitor’s alert email does not reach the Recipients list: monitor alerts go to the addresses in the site’s own Email recipients field, and to its own Slack and JSON webhooks.
Section titled “Related”