The short answer
A WordPress site that keeps going offline is usually running out of something for a few minutes at a time: server memory, CPU, or the number of PHP processes it’s allowed, often because of bot traffic or a scheduled job. First check whether it’s down for everyone or just for you, because a security firewall can block your own connection while the site works for everyone else. Then note the exact times it goes down, because the timing usually points to the cause.
A site that’s simply broken is easier to fix than one that keeps coming and going. By the time you’ve rung your host, it’s back up, and they tell you the server looks fine. They’re often right: at that moment, it is. The problem happens in bursts, so the evidence is in the logs and graphs from those few minutes, not in how the server looks when someone checks it later.
Is this what you’re seeing?
- Your WordPress site keeps going offline, then comes back by itself after a few minutes or an hour.
- The site loads intermittently: sometimes quickly, sometimes slowly, sometimes not at all.
- You see timeouts, “This site can’t be reached”, or Cloudflare error pages such as 521, 522 or 524.
- It goes down at similar times each day or week, or when you send a newsletter or post on social media.
- It’s down for you, but customers or colleagues say it’s working.
- Restarting the server fixes it for a while, then it happens again.
What usually causes it
The server runs out of resources for a while
Every hosting plan has limits: memory, CPU, and how many PHP processes can run at once. When they’re used up, new visitors wait in a queue and eventually get an error. On shared hosting, the host may also slow or pause your account for a few minutes if you go over your share. Small cloud servers have a twist: many of them allow short bursts of extra CPU, and once that allowance is spent, the server slows to a crawl even though its graph says it’s hardly busy. How to tell: your hosting control panel has resource-usage graphs or “faults” figures. Look at them for the times the site went down.
Bots and scanners
Search engine crawlers, AI crawlers and automated vulnerability scanners can send hundreds of requests in a short burst. Each one that isn’t served from a cache makes WordPress do real work. Requests that are slow or never finish can tie up the server’s workers for far longer than a normal visit. How to tell: the access log for the time of the outage shows lots of requests from a handful of addresses, or for files that don’t exist on your site, such as .env. I’ve written about this in my site goes down when bots crawl it.
A firewall blocking you, not everyone
Security plugins and server-level tools ban addresses that look suspicious. Their rules aren’t always well judged, and a few attempts to log in, or a few visits to the login page, can get your own office or home connection banned. To you, the site is down. To everyone else, it’s fine. How to tell: load the site on your phone with Wi-Fi turned off. If it works there and not on your office connection, something is blocking your address.
Scheduled jobs at the same time every day
Backups, security scans, imports, newsletter sends and WordPress’s own scheduled tasks all use the same server as your visitors. A heavy backup or a full malware scan running at the busiest time of day can be enough to push a small server over. How to tell: the outages happen at roughly the same time, and match the schedule of a backup or security plugin.
Something outside the server
Less often, it’s not the server at all: a domain or SSL certificate that has expired or failed to renew, DNS settings that changed, or a problem at Cloudflare or your host that affects many sites at once. How to tell: the error mentions the certificate or says the domain can’t be found, or your host’s status page reports an incident. If it started after a DNS change, see site went offline after changing DNS.
What you can safely try
Take a backup first, ideally at a quiet time so the backup itself doesn’t add to the load. None of these steps change how your site works.
- When it next goes down, check from your phone on mobile data and ask someone elsewhere to try. Down for you only points to a block on your connection. Down for everyone points to the server.
- Write down the date, the time and the exact error each time. Screenshots help. Three or four outages with times are worth more than any amount of guessing.
- Set up a free uptime monitor that checks the site every few minutes and emails you when it fails. It gives you an accurate record of when and how long.
- Look at your hosting control panel’s resource-usage graphs for those times, and ask your host for the error logs covering them. Asking for specific times gets a much better answer than “my site keeps going down”.
- Check when your backup, security-scan and other scheduled tasks run. If one matches the outages, move it to the early hours, or ask whoever set it up to.
- Make sure page caching is switched on and working for logged-out visitors. Cached pages are served without WordPress doing the work, which leaves far more room for real visitors.
When to stop and get help
- Your host says the server is fine, but the outages keep happening.
- You’re losing orders, bookings or enquiries while it’s down.
- The fix on offer is “upgrade to a bigger plan”, and you’re not sure that’s the real problem.
- You run your own server, and restarting it is the only thing that helps.
- The logs show requests for odd files, or you see redirects or pages you didn’t create. If you think it’s been hacked, get a hacked site clean-up rather than tuning the server.
A real example
I’ve written up two of these as case notes, because each one looked like a general hosting problem and turned out to be something more specific.
A business site on a small cloud server was slow and kept dropping out, with Cloudflare showing 524 and 521 errors. The hosting dashboard showed the CPU at only around 12%, which made the server look nearly idle. Inside the server, the picture was the opposite: it was spending most of its time waiting for CPU it wasn’t being given, because its allowance of burst capacity had run out. I moved it to a larger server, tested it before switching over, and the case note records that response times went back to normal. The short version is in when a WordPress site kept going offline, and the full technical write-up is in how I fixed intermittent Cloudflare 524 errors.
On another occasion, just after I’d taken over a site, the owner reported it was down, and from my browser it was. The server itself was healthy. There were real problems worth fixing, including scanner traffic and a web server allowed far more workers than its memory could support. But the reason I couldn’t reach the site was that its own firewall had banned my address after a few login attempts. That’s in the site that looked down even though the server was fine.
In a separate incident on a larger server, a burst of crawler traffic was followed by a chain of failures, and the likely cause was the server running short of memory under too many simultaneous requests. I tell that one in my site goes down when bots crawl it.
Getting it fixed
Send me the times it went down and what you saw, if you have them. I look at the logs and resource figures for those times, work out which layer is failing, and fix that, rather than starting with a bigger hosting plan. A WordPress repair is a fixed £119, which covers finding the problem and up to 90 minutes of fixing it. Most repairs fit in that, and you’re never charged more without agreeing first. Server-level work that needs longer is £75 an hour, agreed with you before I start.
If your site is down and it’s costing you money right now, see emergency WordPress support for a faster start.
Questions
My hosting company can’t explain why my site keeps going offline. What now?
Hosts usually check whether the server is running, and it usually is by the time they look. Give them exact times and ask for the error and resource logs for those minutes. If they still can’t say, it’s worth having someone look across WordPress, the server and the traffic together. That’s the part I do.
Why does my WordPress site load intermittently, fine one minute and slow the next?
Usually because some pages are served from a cache and some aren’t, or because the server is close to a limit and tips over when a few more visitors arrive. A page that’s fast in a private window and slow when you’re logged in is normal. A page that’s slow for everyone at certain times isn’t.
Will a bigger hosting plan fix it?
Sometimes, and on one of the sites above a larger server was the right immediate step. But if the load comes from bots, uncached pages or a badly timed job, a bigger plan only delays the next outage. It’s worth knowing why it’s running out before paying for more.
What do 500, 502 and 503 errors mean when the site keeps dropping out?
They’re different ways of saying the server couldn’t produce the page. A 503 often means the server is too busy, a 502 that the part running PHP didn’t answer properly, and a 500 that something failed while building the page. See 500, 502 and 503 errors in WordPress.
Could my site be going offline because it’s been hacked?
It’s possible but not the usual reason. Seeing scary-looking requests in your logs, for files like .env or odd PHP files, usually just means automated scanners are trying their luck, and the files don’t exist. If you also see redirects, spam pages or warnings from Google, get a hacked site clean-up.
Written by Peter Claridge, who fixes WordPress sites for businesses across the UK from Reading. Last checked on WordPress 7.1, November 2026.