Blog · November 17, 2026 · 12happy team
WordPress uptime monitoring: plugin or external monitor?
Search the plugin directory for “uptime” and you’ll find plenty of answers. Most of them come down to a question that matters more than which plugin to pick: where does the check run?
A check that runs inside your WordPress site can’t tell you that your WordPress site is down. When the server, PHP or the database stops, the code that was supposed to notice stops with it. Uptime monitoring has to come from outside: another machine that requests your pages the way a visitor would, and tells you when the answer is wrong.
This guide covers why that is, which WordPress failures a basic check misses, and how to set up monitoring that catches them. Everything here works on a free plan.
Why a plugin can’t watch the site it runs on
WordPress only runs when something requests a page. That includes its scheduler, WP-Cron, which runs its jobs when someone visits the site. A plugin that “checks your uptime” from inside WordPress has the same limits as the rest of your site:
- If the server is down, nothing runs, so nothing sends the alert.
- If the database is down, WordPress stops before plugins load and shows “Error establishing a database connection”.
- If a plugin update causes a fatal error, the monitoring plugin goes down with everything else.
The plugin would be the last to know, and it can’t tell anyone.
“Uptime plugins” that are really external services
Some plugins do work, because the plugin isn’t doing the checking. It connects your site to a service that checks from its own servers. Jetpack’s downtime monitor is the best-known example. According to Jetpack’s documentation, it’s free and works like this:
- every five minutes, a HEAD request to your home page;
- a failure is confirmed from three servers in other locations before it counts;
- alerts go by email, to the site’s administrators, at their WordPress.com account address. You can’t add other recipients.
That’s a real external monitor, and for a personal blog it may be all you need. Its limits are the ones to know: it watches one URL, it only looks at whether the page answers, and the alert goes to one kind of inbox.
What can go wrong on a WordPress site, and what a monitor sees
Most WordPress outages do produce an error code, which any external check will catch. A few look fine from the outside. The status codes below are WordPress’s own, taken from its source code:
| What happens | Status | Caught? |
|---|---|---|
| Server or host down | none | Yes |
| Database down | 500 | Yes |
| Fatal PHP error | 500 | Yes |
| Update running or failed | 503 | Yes |
| SSL certificate expired | TLS error | Yes |
| Cache serves a saved copy | 200 | No |
| Site hacked or defaced | 200 | No |
| One page broken | 200 | No |
The first five are the familiar ones. A database outage shows “Error establishing a database connection”; a plugin or theme that breaks PHP shows “There has been a critical error on this website”; an update shows “Briefly unavailable for scheduled maintenance”, which WordPress drops on its own after 10 minutes if the update failed. An expired certificate is caught as long as the check verifies it.
The last three are the expensive ones, because they can go on for days. The page answers 200, so a check that only looks at the status code stays green. To catch them, the check has to read the page.
Setting it up: three checks for a WordPress site
The steps below use 12happy, but the idea carries over to any external monitor that supports keyword checks. On the free plan you get 50 monitors checked every 5 minutes, and every failure is confirmed from a second location before you’re alerted.
1. The home page, with a keyword
A keyword check passes only when the page answers and contains a piece of text you choose.
- In the console, choose Add monitor.
- Under What should we watch?, enter your address:
https://yoursite.com. - Under Check type, choose Keyword.
- In Keyword that must be present, enter text that appears only when the page is really working. Good choices: a phrase from your footer, your site’s tagline, or the title of a section near the bottom of the page. Avoid your site name: it also appears in error pages and in many hacked pages.
- Pick How often, and tick the alert channels that should hear about it.
Text near the bottom of the page is the best test. If PHP dies halfway through rendering the page, the top may arrive and the footer won’t.
2. A page your cache can’t answer
Page caching is good for speed, and it’s also why a WordPress site can look up while it’s broken. If your cache plugin or CDN serves the home page from a saved copy, the first check can pass even when PHP and the database aren’t working.
Add a second monitor on a URL that has to go through WordPress every time. The REST API index is a good candidate, because page caches don’t usually store it:
- Address:
https://yoursite.com/wp-json/ - Check type: Keyword
- Keyword:
namespaces(every REST index response includes it)
If that URL returns an error even while the site is working, a security plugin may have turned off the REST API for visitors who aren’t logged in. Use another page that skips the cache instead, such as one your cache plugin’s settings exclude.
3. The page that makes you money
On a WooCommerce shop, a booking site or a lead-generation site, the home page isn’t the page that matters most. Add a keyword monitor for it as well:
- a product page, with the text of your Add to cart button;
- the contact or booking page, with the label of the form’s submit button.
A plugin update that breaks the checkout leaves the home page untouched. This is the check that notices.
Check every 5 minutes or every 30 seconds?
Five minutes is plenty for a blog or a brochure site. For a shop, an outage that lasts four minutes can fall entirely between two checks: your customers saw it and your monitor didn’t. If your site earns money every minute (see what an outage costs you), or you promise a client an uptime figure, check more often. Paid plans check every 30 seconds (Pro) or 15 seconds (Agency). The nines of availability shows which interval can prove which uptime target.
Avoiding false alarms
External monitors look like bots, because they are. If you get alerts while the site loads fine for you, look at these first:
- Security plugins and firewalls that block unknown bots or rate-limit requests can answer a monitor with 403 or 429. The monitor’s activity shows the status code it got. If that’s the cause, relax the bot or rate-limit rule for the URLs you monitor.
- Slow hosting. On a busy shared server, the first request after a quiet spell can take several seconds. If checks time out now and then, raise the timeout under the monitor’s advanced settings rather than ignoring the alerts. If they time out often, that’s worth knowing too.
- Confirmation. A single failed request from one network path isn’t an outage. 12happy re-checks from a second location before it alerts you, on every plan.
Plugin or external monitor: the short answer
| Notices… | Inside WordPress | External monitor |
|---|---|---|
| Server down | No | Yes |
| Database or PHP error | No | Yes |
| Cached or hacked page | No | With a keyword |
| Pages beyond the home page | Varies | Any you add |
An external monitor can also reach you where you’ll see it. On 12happy that’s email, Slack, Telegram and push on every plan, with SMS and webhooks on Pro.
Use an external monitor, and add keyword checks for the pages that matter. If you already run Jetpack, keep its monitor on: a second, independent check costs nothing. If you look after WordPress sites for clients, 12happy Agency gives each client a workspace and a status page of its own, and every public status page gets live status badges for the client’s site.