Building MetricPoints: Know a Site Is Broken Before the Client Does

MetricPoints started as monitoring and security tooling: uptime, security headers, Content Security Policy, and frontend errors. I am rebuilding it around a more specific job.

Help agencies know a client website is broken before the client does.

That shift is documented on the MetricPoints blog. This article is the version I want on my own site, because the product is also a statement about the kind of production work I take seriously.

The Failure I Keep Getting Called About

I work with agencies that run WordPress sites, and some that run custom sites alongside them. The teams are usually busy. Design, content, hosting questions, contractors, proposals, and support already fill the week.

The failure that damages the relationship is rarely a dramatic outage announcement. It is quieter:

• a form stops submitting
• checkout cannot finish
• a plugin update changes something important
• a third-party script breaks the page
• the site is slow enough that it might as well be down
• several sites on shared infrastructure fail together, and each one looks like a separate incident

Sometimes the client finds it. Sometimes the client's customer finds it first.

Uptime still matters. So do SSL, DNS, domain expiration, security headers, browser errors, and Content Security Policy reports. A site can return a successful response and still fail at the thing the business needed it to do. A lead form that does not submit is not a healthy site.

What I Am Focused On

MetricPoints is becoming an agency website-assurance layer. The working view I want is a portfolio, not another inbox.

For each client site, an agency should be able to see:

• what is critical
• what needs review
• what is only advisory
• where monitoring is still incomplete
• where the evidence lives if someone has to investigate

I do not want a green checkmark with nothing behind it. If there is not enough evidence to call a site healthy, the honest status is that monitoring is incomplete. That is more useful than false confidence.

I also do not want the product to begin with an installation. Strong external monitoring comes first: availability, certificates, DNS, and the public behavior of the site. Browser script and the WordPress plugin add context when the agency is ready for them. Protection should be useful before anyone has to put code on the site.

The other focus is noise. Agencies already live inside ManageWP, host dashboards, plugin notices, uptime emails, and security scans. Another firehose does not help. The product has to sort signal from evidence, so the thing that needs attention is distinguishable from the thing that is only available if someone goes looking.

Why This Is Harder Than an Uptime Check

A portfolio view has to connect problems that do not look connected from a single site.

Shared hosting trouble, a bad plugin release, bot traffic, a broken tag manager, and a DNS change can all show up as "the site is being weird." MetricPoints should help an agency see when several client sites are failing for a related reason, and it should keep the evidence attached to that conclusion.

It should also stay calm. Critical, review, advisory, and incomplete are different kinds of information. Mixing them together is how important failures get buried.

How This Connects to the Rest of My Work

MetricPoints is the productized version of work I already do for clients: production visibility, CSP and frontend diagnostics, and practical investigation when a site is up but not actually working.

The Website Reliability offering on this site is the managed version of that oversight. MetricPoints is the tool underneath it, and it is also the proof that I build this kind of system instead of only recommending one.

The same judgment shows up in custom work. Dashboards, portals, and WordPress plugins are only useful if someone can tell when they fail in production. Monitoring that stops at "the server responded" is not enough for a business that depends on a form, a checkout, or a workflow.

The Standard I Am Building Toward

Know before the client knows. Then have the evidence to do something about it.

That is the focus. Not a broader security suite. Not a louder alert stream. A clear view of the sites an agency is responsible for, with enough honesty to say when the picture is still incomplete.

The longer essays live on the MetricPoints blog, including What I'm Building With MetricPoints and the pieces on agency website assurance. This site is where I connect that product back to custom development, WordPress work, and ongoing reliability.