Blog / How the internet works /
Why websites go down: the most common causes
A website going down always feels sudden, even when it was a long time coming. One minute it’s fine; the next someone is sending you a screenshot of an error page and asking if you’ve seen it.
Part of my job as a web administrator at Royal Arts Academy is keeping a site up and working out what happened when it isn’t. What follows is the general picture, not a story about one incident: the causes that account for most downtime, in roughly the order I’d check them.
First, work out what kind of “down” it is
“The site is down” can mean several different things, and they point to different causes. Before you touch anything, answer these:
- Is it down for everyone, or just you? Try another device, mobile data instead of Wi-Fi, or a free “is it down” checker.
- What does the browser say? A timeout, a “server not found”, a security warning and a white page with an error number are four different problems.
- Is the whole site down, or one page?
- What changed recently? An update, a new plugin, a DNS change, a renewal date.
That last question solves a surprising number of cases.
The common causes
1. The domain expired
Boring, embarrassing, and common. If the domain isn’t renewed, it stops working, and in some cases the registrar parks it on an advert page. Everything goes: the site and the email.
Check: look up the domain’s expiry date with your registrar. Prevent: turn on auto-renew, keep a payment method current, and put the renewal date in a shared calendar. The registrar’s emails should go to an address that someone actually reads.
2. DNS problems
If the domain points nowhere, or to the wrong place, visitors can’t find the site. This often follows a change: a host move, new nameservers, a deleted record. I explained the moving parts in DNS explained for people who just want their site to work.
Check: a DNS lookup tool shows what the world sees. Prevent: keep a copy of your records before any change.
3. The hosting account has a problem
Hosting that wasn’t paid, a plan that’s been suspended for going over its limits, or a server that’s having its own bad day. Shared hosting can also slow or suspend an account if your site uses too many resources.
Check: your host’s status page and your account dashboard. Prevent: keep billing current, know your plan’s limits, and have a way to contact support quickly.
4. The server ran out of something
Full disk. Out of memory. Too many simultaneous visitors. A site on a small plan can fall over when a post gets shared widely or a bot hammers it. Typical signs are 500 or 503 errors, or a very slow site that eventually gives up.
Check: resource graphs in your host’s panel, and error logs. Prevent: caching, a plan that fits the traffic, and clearing out old backups and logs that quietly eat disk space.
5. A bad update or plugin conflict
On WordPress this is the classic. A plugin or theme update, or a new plugin, clashes with something else and the site shows a white screen or a critical error message. A PHP version change on the host can trigger the same thing when old code doesn’t cope with a newer version.
Check: what changed just before it broke. Turn off the latest plugin through your hosting file manager if you can’t reach the dashboard. Prevent: back up before updating, update one thing at a time, and test big updates on a staging copy where possible.
6. The SSL certificate expired
The site is technically up, but browsers put up a full-page warning and most visitors leave. See HTTPS and the padlock for what the warnings mean.
Check: click on the warning to see the certificate dates. Prevent: make sure auto-renewal works and re-check it after any move.
7. The site was hacked or infected
Compromised sites can be defaced, redirected to spam, or taken offline by the host once malware is detected. This is a bigger problem than downtime, and speed matters.
Check: scan with a reputable security tool, check Google Search Console for security warnings, and look for files or users you don’t recognise. Prevent: updates, strong passwords, two-factor login, removing unused plugins and accounts.
8. A configuration mistake
A wrong line in a configuration file. A redirect loop. A database password changed on the server but not in the site’s settings. These are usually self-inflicted and usually follow somebody “just quickly changing” something.
Check: the error message, which is often clearer than you’d expect, and the change log if you keep one. Prevent: keep notes of what was changed and when.
9. Something outside your control
A wider outage at your host, a data centre problem, or a major network incident. Not frequent, but real. You can’t fix it, but you can communicate about it.
Check: your host’s status page. Prevent: this is where a good host, backups elsewhere and a plan for talking to users earn their keep.
A quick triage order
When it’s down and you want to know where to start, this order saves time:
- Confirm it’s down for others, not just you.
- Check the domain is active and hasn’t expired.
- Check DNS resolves to the right place.
- Check the host’s status and your account.
- Look at the error: certificate, server error, or a white screen.
- Think about the latest change, and undo it if you can.
- Restore from a backup if you must.
Most causes show up by step four or five.
How to find out before your visitors do
An external uptime monitor pings the site every few minutes from outside and alerts you by email or message when it fails. Many free options exist. Set it to check for a word on the page, not just a response, so it catches the sites that load but show an error. Send the alert to someone who can act.
What to do afterwards
Treat every outage as free information.
- Write down what happened, when, and what fixed it.
- Ask what would have caught it earlier.
- Fix the process, not just the symptom. If an expired domain was the cause, the fix is a calendar reminder and a shared billing contact, not an apology.
Key takeaways
- “Down” has different meanings; check whether it’s down for everyone first.
- Common causes: expired domain, DNS, hosting account, server limits, bad updates, expired certificates, hacking, and configuration mistakes.
- Always ask what changed recently.
- Back up before changes, monitor from outside, and keep renewals visible.
FAQ
How long should I wait before panicking?
Not long enough to be sure it’s just a blip? Check from another network first, then work through the triage order. Waiting rarely fixes expired domains or full disks.
Does downtime hurt SEO?
Short outages usually cause little lasting harm. Repeated or long ones can mean Google reduces how often it crawls your site, and visitors lost in the meantime don’t always come back.
Should I have a backup host?
For most small sites, good backups stored elsewhere are enough. A second host ready to go matters for sites where every hour offline has a clear cost.
Related reading
- DNS explained for people who just want their site to work
- What is a CDN, and do you need one?
- What actually happens when you type a web address
What is the one renewal date for your site, whether domain, hosting or certificate, that you couldn’t name from memory right now?