Quick start

From nothing to a website that is genuinely being watched — about five minutes, most of it spent deciding who should be told when it breaks.

On this page
  1. Create a workspace
  2. Point it at a website
  3. What happens from then on
  4. Decide who hears about it
  5. Bring in the rest

Create a workspace#

Signing up takes an agency name, your name, an address and a password. No card is asked for and none is stored. You get 14 days of full access; nothing is charged when they end, because there is nothing on file to charge — an invoice is raised and you decide what to do with it.

The workspace is created with one plan already in it, called Standard (set your price). That is the plan you sell to your clients, and it is priced at zero on purpose: we do not know what you charge, and a number invented here would end up on your first invoice.

The address you sign up with is not confirmed, because this installation has no mail transport configured. Nothing was sent to it and you are not waiting for anything. It matters later — see Alerts.

Point it at a website#

The first screen asks for one thing: an address. Paste it, and the console creates the client, attaches the plan, adds the site and runs a check immediately — the status usually lands within a second or two.

Behind that one field it also runs a first sweep of the slower checks: the TLS certificate, the domain registration, the security headers, and the DNS records it will compare against from then on. You do not have to ask for any of them.

Everything it guessed is editable. The client is named after the hostname unless you said otherwise, and renaming it is one field on the client page.

What happens from then on#

The site is checked on its interval — every 60 seconds on the starting plan, and never faster than every 30 seconds, which is how often the monitoring loop wakes up. Each check records the status code, how long the first byte took and how big the response was.

When a check fails, an incident opens by itself, carrying the reason and the last status code. When the site recovers, the incident closes and the downtime window is written into the uptime figures. You never open or close one by hand unless you want to acknowledge it.

Decide who hears about it#

Until you add a channel, an outage is recorded and nobody is told. Add one on the Alerts page: an e-mail address, a Telegram chat, a Slack or Discord channel, PagerDuty, or your own webhook.

On this installation e-mail and Telegram are not configured, so a channel of either kind is saved and its deliveries are recorded as skipped rather than sent. That is deliberate — the alternative is a product that says it told you. Webhooks are unaffected and work today.

Bring in the rest#

Add the other sites one at a time, or import a spreadsheet — the import page takes a CSV of client, name and URL. Each client can then be given a login to their own portal, which shows their sites, their uptime, their incidents and their invoices, and nothing belonging to anybody else.

When you are ready to charge for it, see Billing your clients: set a price on the plan, assign subscriptions, and generate invoices from them.