How I Watch 50+ Client Sites From One Tab

I look after about 50 client sites at any given time. WordPress installs, Shopify stores, a few custom Laravel apps, a couple of static things on Netlify. Different hosts, different CDNs, different SSL setups. Some run cron jobs. Some have APIs. One has a WP-Cron so temperamental it breaks every other Tuesday for reasons I have genuinely never worked out.

Here's how I keep all of them honest without drowning in dashboards or finding out about an outage from the client.

The baseline: what every site gets

I don't sit and configure monitoring per site. There's a standard set of checks every new client gets, no exceptions:

Uptime check: an HTTP request every 2 minutes, re-confirmed from a second region. I set the expected status code (usually 200, sometimes 301 for redirect domains) and I'm done. The dual-region part is the whole point: I only get woken when the site is actually down, not when one probe and one server had a bad ten seconds between them.

SSL monitoring: it watches the cert and warns at 30, 14 and 7 days out. After the Saturday incident that nearly cost me a client, I don't trust auto-renewal to just work. Not without a net under it.

DNS monitoring: watches A, MX and TXT records. You'd be amazed how often records change with nobody meaning them to. A domain transfer that quietly drops the MX records will break a client's email for days before anyone notices. This is the check most people skip, and it's the one that saves you when something genuinely weird happens.

Setup time per site: about 2 minutes.

Premium clients get more

For clients on bigger plans, or ones with hairier setups, I bolt on a few extras:

Broken-link detection: a weekly crawl that finds dead links. Worth it for shops where products come and go, and for content sites that link out to things that move or vanish. I don't put it on every site. On a site that barely changes it just generates noise.

Lighthouse audits: weekly scoring for performance, SEO and accessibility. This one isn't about catching emergencies. It's proof of work. When a client asks what you've actually done for them lately, "your Lighthouse performance score went from 62 to 89 after we cleaned up the images" is a hard number they can't argue with.

Cron monitoring, for sites with scheduled jobs: backups, imports, email digests. The job pings Hoppla once it finishes cleanly. No ping by the deadline, I get an alert. Dead simple, and it catches the silent failures you'd otherwise only find out about when someone asks why the weekly report stopped a month ago.

GEO audit. This one's newer. It scores how ready the site is for AI search engines: structured data, whether you're blocking AI crawlers, how citable the content is. I switch it on for clients who care about SEO, because AI search is turning into a real way people find sites and most of them aren't set up for it.

The alert setup that actually holds up

The checks are half the system. The other half (honestly the more important half) is how the alerts get routed.

I learned this the hard way: spread alerts across channels and some will get missed. Doesn't matter how disciplined the team is. If uptime goes to #monitoring-uptime and SSL goes to #monitoring-ssl, someone mutes one of them during a hellish week and misses the single alert that mattered.

So everything lands in one Slack channel. One. Every check type, every client. It's called #monitoring and it's the only one anyone has to watch.

Escalation is a flat rule: page me right away. No ack inside 15 minutes, page the backup. Silence at 30, it calls a phone. That covers nights, weekends and holidays without a fiddly on-call rotation, though we run one of those too now that the team's grown.

The other setting I'd die on a hill for: dual-region confirm on every uptime check. Before I turned it on I was eating 2-3 false alarms a week. Enough that I started half-ignoring the phone, which is the worst place you can end up. With the second-region recheck, false alarms went to basically zero. Now every alert is a real one.

Status pages: the thing that saves the most time

For about 20 of my clients (the premium ones) I run branded status pages on their own domains (status.clientname.com).

I set them up because clients kept asking and I figured it was a nice-to-have. Wrong. It's the single biggest time-saver in the whole setup.

Here's the why. When something goes down, the status page updates itself. The client, and their customers, can look there instead of emailing me. I don't have to type "we're aware and working on it" with one hand while fixing the actual problem with the other. The page says it for me.

The "is the site down?" emails dropped by roughly 80% after I rolled status pages out. Not a figure of speech. I counted.

And at renewal, I can put 12 months of uptime in front of the client. "99.97% uptime, incidents acknowledged in about 2 minutes" is a very hard thing to walk away from.

Making it pay for itself

I bill monitoring as its own line item. Not a fortune, €25-35/month for the standard package (uptime + SSL + DNS + status page), more for the premium one with Lighthouse, broken links and cron.

Most clients say yes without blinking. They know monitoring costs something, and they'd much rather hand over a small amount than check their own sites by hand and hope.

The maths is kind: the platform costs me a flat monthly fee no matter how many sites I watch. So every client on a package is almost pure margin. 20 clients at €30/month is €600/month for something that basically runs itself.

I didn't set out to turn monitoring into a profit centre. But solve a real problem well and people are happy to pay for it, even when it costs you almost nothing at the margin.

The 30-minute monthly review

Once a month I give the dashboard 30 minutes:

• Any new client sites that still aren't being monitored?
• Any checks crying wolf (threshold needs a nudge)?
• Lighthouse trends, anyone sliding?
• SSL certs due in the next 45 days?
• Domains expiring in the next 90?

This is what catches the slow-burn stuff that never trips an alert. A Lighthouse score that's drifted down for three months. A domain that lapses in 60 days with nobody on it. A broken-link count creeping up.

30 minutes a month. That's the whole time cost of monitoring 50+ sites once it's set up.

If you look after client sites and your setup is messier than this (or you don't have one), try this approach. Setup runs about 2 minutes a site. The first time it catches something before your client does, you'll wonder what took you so long.