Do you monitor my sites or only the server

We monitor the infrastructure: the machines, the network, the services that make hosting work. That tells us when a server is in trouble. It does not tell us when your particular site returns an error, because from the outside that looks like a working server serving a page.

The distinction that matters

Server monitoring answers "is the platform healthy". Site monitoring answers "is my site doing what it should". They fail in different ways:

• A server can be perfectly healthy while your site returns a fatal error after a plugin update.
• A site can load fine while its checkout has been broken for a week.
• A site can be up and serving a defaced page.

None of those trigger an infrastructure alert, because nothing about the infrastructure is wrong.

What you should watch yourself

On an estate this is not optional, because you will not notice a quiet failure on the eleventh site by browsing. At minimum:

• Availability per site, not per server. An external checker hitting the actual URL.
• The status code, not just whether something responded. A 200 serving an error page is the case that fools simple checks.
• A string on the page. Checking that expected content is present catches defacement and blank renders that a status code misses.
• Certificate expiry, which is a scheduled failure you can prevent entirely.

Where the boundary sits in practice

If the cause is on our side, we are already working on it before a ticket arrives. If the cause is in your site's code, its database, or a change someone made, we will help you find it but nothing alerted us, because nothing alerted anyone.

Telling those two apart quickly is a skill worth having, and it is covered in telling a host problem from a site problem.

When something is down

Work the first hour deliberately rather than opening a ticket and waiting. Most site level outages are diagnosable from the error log within minutes, and arriving with the actual error gets a real answer instead of a request for more detail.

The sequence is in the first hour triage checklist, and what to bring to a ticket is in collecting the right evidence.

The practical arrangement

Let us watch the platform, and put a monitor on your own sites. The two together cover the ground; either alone leaves a category of failure nobody sees. What the uptime figure itself covers is explained in what the 99.9% figure actually covers.