Limits on one domain

We limit how hard we check any one domain, so that nobody can point our checks at a server that is not theirs and flood it. When a new Monitor would go past one of these limits, the dashboard refuses to create it.

The limits never stand in the way of watching your own domain: on every Plan, all the Monitors your Plan allows fit on one domain, at any interval.

What counts as one domain

A Monitor counts against the domain its host is registered under, not the full host: shop.acme.com and api.acme.com:8443 both count as acme.com, and so does a Server Monitor of acme.com. Website, Port and Server Monitors count together, because the same server receives them all. A Monitor whose host is an IP address counts against that address.

On a platform that gives each customer a name under its own domain, each name counts on its own: a.vercel.app and b.vercel.app are two domains.

The limits

Each limit is counted in each location on its own, because every location checks your Monitors on its own schedule. In any one location, a domain receives at most:

• 500 Monitors of yours.
• 1,000 requests a minute from your Monitors.
• 3,000 requests a minute from everyone’s Monitors, yours included.
• 1,000 MB a minute read by your Monitors with assertions. Each of their requests counts as the most we may read, 1 MB, whatever it read.

A Monitor sends one request each interval from each of its locations: at the fastest interval, every 30 seconds, that is 2 requests a minute. A paused Monitor counts, because it will resume. A Suspended one does not. Second tries, other locations checking to confirm a failure, and the Test before you create a Monitor are not counted.

When a Monitor is refused

The refusal shows under Locations, and names the domain and the location where the limit was reached. To create the Monitor, do what it says, which is one of these:

• remove a Monitor on that domain,
• choose a longer interval, so each Monitor sends fewer requests,
• choose other locations, since each location counts on its own.

The limits are checked only when you create a Monitor. Once it exists, it is always checked: we never skip a check to stay under a limit, because a skipped check would look like an outage.