How to monitor WordPress uptime, and trust the alerts you get
A WordPress site can be down for ten minutes before anyone tells you. This guide covers what to check, how often, from how many locations and how to read the numbers, whether you run one site or fifty.
The short answer
To monitor WordPress uptime, use an outside monitor that requests your site about every 60 seconds, checks the page content as well as the HTTP status, runs from at least three locations and only alerts you when several locations agree the site is down. Add SSL certificate and domain expiry checks, and track the result as a monthly uptime percentage: 99.9 percent allows about 43 minutes of downtime in a 30-day month.
What is WordPress uptime monitoring?
WordPress uptime monitoring is an outside service that requests your site on a schedule, checks that it answers correctly and quickly, and alerts you when it does not. It also keeps a record, so you can say how many minutes a site was down last month and why.
The phrase "outside service" matters. A monitor has to sit outside your hosting, because anything running inside WordPress goes down with it. If the server, PHP or the database fails, a plugin cannot tell you. A checker on another network can. This is the core of website uptime monitoring for any site, but WordPress adds its own failure modes, such as plugin updates, PHP limits and database connections, which is why this guide focuses on it.
Uptime monitoring is not the same as WordPress performance monitoring. Uptime asks whether the site is reachable and serving the right page. Performance asks how fast it loads. You need both, but uptime comes first, because a fast site that is down earns nothing. If you run more than one site, it also helps to manage multiple WordPress sites from one dashboard, so every site's uptime sits in one place.
What a WordPress uptime monitor should check
A check that only reads the HTTP status code will call a broken site healthy. These are the checks worth having, in order of value:
- Status code and response time. Anything other than a successful response is a failure, and a rising response time is an early warning.
- Expected content on the page. This catches the WordPress white screen of death, a database error page and a maintenance page that never went away, all of which can return 200 OK.
- The URLs that earn money. Watch a checkout, a booking form or a landing page, not only the home page.
- SSL certificate expiry. An expired certificate shows visitors a browser warning that most will not click through, which is an outage in everything but name.
- Domain expiry and DNS changes. A lapsed domain or an unexpected DNS change takes a site offline, and no server monitoring will notice.
- Pages behind a login. Useful for account areas and membership sites that anonymous visitors never see.
WPCentrify covers all of these in its WordPress uptime monitoring feature.
How often should you check a WordPress site?
The check interval sets how long an outage can run before anything notices it. Most good monitors also wait for several failed checks in a row before alerting, to avoid reacting to a single blip. Put those two together and the slow intervals get slow quickly:
| Check interval | Wait before first failed check | Worst case to confirm |
|---|---|---|
| 1 minute | Up to 1 minute | About 3 minutes |
| 5 minutes | Up to 5 minutes | About 15 minutes |
| 15 minutes | Up to 15 minutes | About 45 minutes |
Worst-case figures for any monitor that needs three failed checks in a row to confirm an outage. A 99.9 percent target allows only about 43 minutes of downtime in a 30-day month.
A 15 minute interval can take 45 minutes to confirm an outage, which is more than a whole month's downtime budget at 99.9 percent. That is why 60 seconds is the practical target for a site that matters. Many free monitors check every five minutes from a single location, which is fine for a hobby site and thin for a client's business.
Why one location causes false alarms
A monitor that checks from one place cannot tell a real outage from a network problem near itself. If the checker loses its route to your server, it reports the site as down even though every real visitor can still reach it. After a few of those 3am alerts, people start ignoring alerts, and that is when a real outage gets missed.
The fix is agreement. Check from several locations, and alert only when more than one sees the failure. Two details make this work well:
- Use at least three locations. With three, you can require two to agree and still tolerate one checker having a bad day.
- Spread them across providers. Two locations on the same network provider can share the same bad day. Locations on different providers rarely fail together.
Locations also tell you something useful by themselves. If the site is down in Asia but up in North America, you have a regional problem and a very different phone call to make. WPCentrify works this way: WordPress uptime monitoring from three locations checks every site from Malaysia, Canada and the USA, and emails you only when locations at two different providers agree.
What to look for in website uptime monitoring tools
Whether you pick a standalone uptime monitoring service or a WordPress management platform, the same questions separate a useful monitor from a noisy one:
| Ask | Why it matters |
|---|---|
| Check interval of 60 seconds or less? | The interval sets how long an outage runs before the first failed check. |
| Several locations, with agreement before alerting? | One location cannot tell an outage from a network problem near the checker. |
| Content checks, not only the status code? | A broken WordPress site can return 200 OK while showing an error or a blank page. |
| SSL and domain expiry alerts? | Expired certificates and domains cause outages that page checks only see once it is too late. |
| A record of every incident? | You need start, end, duration and cause to answer a client, and to track downtime against a target. |
| Alerts you will actually read? | One alert per confirmed outage beats one per failed check. Noise trains people to ignore alerts. |
| Reporting and a portfolio view? | If you run many sites, you need the worst problems first and a monthly uptime figure per site. |
What about free uptime monitoring? Free plans are fine for a personal site. For a business or client site, check the limits: free tiers commonly have longer check intervals, fewer monitors and one check location, which means slower detection and more false alarms. Work out what an hour of downtime costs before deciding a free tool is enough. WPCentrify uptime monitoring is free during early access.
How to set up WordPress uptime monitoring, step by step
These steps work with any decent monitoring tool, including WPCentrify.
- List the URLs that matter. Start with the home page, then add the pages that earn money or collect enquiries: a checkout, a booking page, a contact form, a key landing page. A home page can load while the pages behind it are broken.
- Choose a short check interval. Pick 60 seconds if your tool offers it. The interval sets how long an outage can run before the first failed check, so it sets how fast you find out.
- Check the content, not only the status code. Choose a phrase that only appears when the page works, such as text in the footer. A WordPress site can return 200 OK while showing a database error or a white screen.
- Check from several locations and require agreement. Use at least three locations, and alert only when two or more see the failure. This is the single biggest fix for false alarms.
- Turn on SSL, domain expiry and DNS alerts. An expired certificate or domain takes a site offline as completely as a server crash, and neither shows up in a page check until it is too late.
- Decide who gets the alert and what they do first. Send alerts to a person who will see them quickly, and write down the first three things that person checks. At 3am, a short checklist beats memory.
- Test the alert. Temporarily change the expected phrase to one that is not on the page and confirm the alert arrives, then change it back. An untested alert is a guess.
- Set maintenance windows and review monthly. Mark planned work so it does not trigger alerts or count against uptime unfairly, then review each site's uptime percentage and incidents once a month.
What is a good uptime percentage?
99.9 percent is the usual target for a WordPress business site: about 43 minutes of downtime in a 30-day month. The differences between targets look tiny as percentages and are large in minutes:
| Uptime | Downtime per 30-day month | Downtime per year |
|---|---|---|
| 99% | 7 h 12 min | 3 days 15 h 36 min |
| 99.5% | 3 h 36 min | 1 day 19 h 48 min |
| 99.9% | 43 min 12 s | 8 h 45 min 36 s |
| 99.95% | 21 min 36 s | 4 h 22 min 48 s |
| 99.99% | 4 min 19 s | 52 min 34 s |
Calculated on a 30-day month of 43,200 minutes and a 365-day year.
Promise only what your hosting can deliver, and measure it with a monitor that checks often enough to see the difference. A five minute check cannot tell you whether a site stayed inside a 43 minute budget. If you look after client sites, a monthly uptime figure belongs in your WordPress client reporting, because clients cannot see the work that prevented an outage.
Track downtime as you go. A website downtime tracker is simply a record of every incident: start time, end time, duration, the error returned and the likely cause. Total it each month against the target, and you can answer "how much downtime did we have?" in one line instead of reconstructing it from memory.
What actually takes WordPress sites down
Most WordPress downtime is dull, which is good news because dull problems have routine fixes. The common causes, and the check that catches each:
- Expired SSL certificate or domain. Caught by expiry alerts, weeks in advance.
- Host or server outage. Caught by the status check, confirmed by several locations.
- Database connection failure. Caught by content checks, because the page shows an error message.
- Memory limits and traffic spikes. Caught by response time creeping up, then by 503 errors.
- A bad update. A plugin or theme update can break a site within seconds. See our guide to a plugin update that broke the site.
When an alert does arrive, our WordPress site down checklist works through the causes in order, and the guide to the WordPress critical error covers the PHP fatal behind many of them.
How to check if a website is up right now
A quick manual check is useful, but it is a snapshot, not monitoring. To check whether a WordPress site is up this minute:
- Load it from a network you are not on, such as mobile data, or ask someone elsewhere to try.
- Run it through a third-party website uptime checker and read the status code, not only the page.
- If it fails everywhere, work through DNS, SSL, hosting, the database and recent changes in that order. Our WordPress site down checklist does exactly that.
Then set up proper monitoring, so you are not the checker.
Monitoring uptime across many WordPress sites
Watching one site is easy. Watching forty is a different job, because the problem is no longer detection, it is noise. Three habits keep it manageable:
- See the worst problems first. A portfolio view that ranks sites by risk beats an alphabetical list. The WordPress maintenance dashboard in WPCentrify opens this way, and uptime monitoring feeds it for every site.
- Alert on confirmed outages only. One email per confirmed outage, not one per failed check.
- Measure each site against its own promise. Different clients get different uptime targets, so track each site against its own.
Uptime is one part of running a portfolio. If you want the whole picture, start with how to manage multiple WordPress sites from one dashboard, where updates, backups, security and client reports sit next to uptime.
How WPCentrify monitors WordPress uptime
WPCentrify checks every connected WordPress site every 60 seconds from Malaysia, Canada and the USA. A site is reported down only when locations at two different providers agree, and only after three failed checks in a row. It is marked up again after two successful checks. You get one email per confirmed outage, with the time it started, the error returned and which locations saw it. SSL expiry, domain expiry and DNS changes are monitored too, and each outage is recorded against that site's uptime figure.
Uptime monitoring is included on every plan, available now in early access, and free during early access. See how WordPress uptime monitoring works in WPCentrify.
Questions about WordPress uptime monitoring
Use an outside monitor that requests your site every minute or so, checks that the right content comes back, runs from several locations and alerts you only when those locations agree the site is down. Add SSL and domain expiry checks, and keep a monthly uptime record for each site.
Every 60 seconds is a good target for a business site. A five minute interval can take up to 15 minutes to confirm an outage when three failed checks in a row are required, which is a third of a 99.9 percent monthly downtime budget of about 43 minutes.
At least three, and the alert should fire only when two or more agree. One location cannot tell a real outage from a network problem near the checker. Locations on different providers are better still, because two locations on one provider can share the same bad network day.
Not on its own. A plugin runs inside the site, so when the server, PHP or the database is down it cannot report anything. A plugin can add useful detail, but the check that decides whether your site is up has to come from outside WordPress.
99.9 percent is the usual target for business sites and care plans. It allows about 43 minutes of downtime in a 30-day month and about 8 hours 46 minutes a year. 99 percent sounds close but allows over seven hours of downtime every month.
Usually because the monitor checks from a single location, so a routing problem or a brief network fault near the checker looks like an outage. Requiring several locations to agree, and several failed checks in a row, removes most false alarms without hiding real outages.
No, not in any way you would measure. A check is one ordinary page request. Checking every 60 seconds from three locations is 4,320 requests a day, which is a tiny fraction of the traffic most business sites already receive.
Uptime monitoring answers whether the site is reachable and serving the right page. Performance monitoring answers how fast it loads and how it scores on measures such as Core Web Vitals. A slow site can be fully up, and a fast site can be down, so you need both.
Look for a check interval of 60 seconds or less, several check locations with agreement before alerting, content checks rather than status codes alone, SSL and domain expiry alerts, a record of every incident, and reporting you can show a client. If you run many sites, also look for a portfolio view and per-site uptime targets.
For a personal site, often yes. For a business or client site, check the limits first. Free plans commonly have longer check intervals, fewer monitors and a single location, which means slower detection and more false alarms. WPCentrify uptime monitoring is free during early access.
Keep a record of every incident with its start time, end time, duration, the error returned and the likely cause, then total it each month against your uptime target. A good monitor does this automatically, so you are not rebuilding the timeline from memory when a client asks.
Request it from a network you are not on, such as mobile data or a third-party checker, and look at the status code and the page itself. If it loads elsewhere but not for you, the problem is local. If it fails everywhere, work through DNS, SSL, hosting, database and recent changes in that order.
Related reading
WordPress uptime monitoring
Three locations, 60 second checks and alerts only when locations agree.
See the featureWordPress site down checklist
Work out why a site is down, in the right order.
Read the checklistManage multiple WordPress sites
Run updates, backups, uptime and reports for every site from one dashboard.
See how it worksKnow before the client calls
Uptime checks from three locations on every site you manage, with one email when an outage is confirmed.
Free during early access. No credit card required.