A website that keeps going down is usually hitting a limit, not breaking at random. The most common causes are a hosting plan that runs out of resources, bots flooding the site, a misbehaving plugin or scheduled jobs piling up at the same time. The fastest way to find yours is to note when the outages happen, then match that pattern to the causes below.
Key takeaways
- Intermittent outages almost always follow a pattern. Write down the date, time and error each time.
- Errors such as 503 and 508 usually mean the site ran out of server resources.
- Bots, including login attacks and aggressive crawlers, can take down a small site on their own.
- Outages at the same time each day point to backups or scheduled tasks.
- Uptime monitoring gives you the timeline you need to find the cause.
Where do I start when a site keeps going offline?
If the site is down right now, follow our guide to what to do when your website is down first. This article is for the frustrating version: the site falls over, comes back, then does it again next week.
Start a simple log. For each outage, note the date, time, how long it lasted and the exact error message. An uptime monitoring service records this automatically, checking your site every minute or so. Two weeks of data usually reveals the pattern.
What does the pattern tell me?
What are the 9 most common causes?
These are roughly in the order we see them on small business sites.
1. The hosting plan runs out of resources
Every plan has limits on CPU, memory and the number of requests it can handle at once. Go over them and the server starts refusing visitors. MDN notes that a 503 error often means a server has hit thresholds like memory, CPU or connection limits (MDN). On many shared hosts, a “508 Resource Limit Is Reached” error means the account has hit its limit on simultaneous connections (LiteSpeed).
Fix: turn on page caching so most visits don’t need PHP at all, then move to a plan with clear CPU and memory allocations if you’re still hitting the ceiling.
2. Bots are flooding the site
Not all traffic is human. Wordfence blocked more than 55 billion password attacks on WordPress sites in 2024 (Wordfence). Each login attempt makes your server do real work. Aggressive scrapers and AI crawlers add more load.
Fix: a firewall that blocks bad bots before they reach WordPress, plus login limits. Our guide to whether you should block AI crawlers covers the crawler side.
3. A plugin or theme is misbehaving
One poorly written plugin can run slow database queries on every page or leak memory until the server gives up. If outages started after an update or a new plugin, this is the prime suspect.
Fix: check the PHP error log for the plugin named in repeated errors. Roll back or replace it, and test updates on a staging copy before they go live.
4. Scheduled jobs pile up
Backups, malware scans, email campaigns and WordPress’s own scheduled tasks all need resources. WordPress’s built-in scheduler only runs when someone loads a page, checking its task list on every page load (WordPress Developer Resources). On a busy site, that can trigger heavy jobs at the worst moment.
Fix: move backups and scans to quiet hours, and ask your host to run WordPress tasks from a real server schedule.
5. The database is struggling
A database bloated with years of revisions, expired temporary data and old logs gets slower until it starts timing out. You’ll often see “Error establishing a database connection” during busy periods.
Fix: clean up and optimise the database, and see our guide to the error establishing a database connection message.
6. The server itself is overloaded
On cheap shared hosting, hundreds of sites can share one server. A busy neighbour can slow yours down even when your own site is fine. If your host’s status page shows frequent incidents, the problem may not be your website at all.
Fix: ask your host for resource graphs and incident history. If they can’t explain the outages, it may be time to move.
7. DNS is unreliable
If some people can reach the site while others can’t, look at DNS. Problems include a slow or failing nameserver, conflicting records or a recent change still spreading.
Fix: check that all your nameservers respond and return the same records. Use a reputable DNS provider with servers in several locations.
8. The site has been hacked
Malware can quietly use your server for sending spam or attacking other sites. That burns through resources and can get the account suspended.
Fix: run a server-side malware scan and check for unknown admin users and files. If you find anything, clean it properly and close the hole that let it in.
9. It isn’t actually down
Sometimes the outage is on the visitor’s side. A security plugin may have blocked your office’s IP address, or one monitoring location may have a network hiccup.
Fix: check from a phone on mobile data. Set your monitor to confirm an outage from more than one location before it alerts you.
How much is this costing me?
Short outages add up. Some customers who hit an error page won’t come back, and Google says server errors make its crawlers slow down, with pages that keep failing eventually dropped from search (Google Search Central). Our article on website downtime cost shows how to put a dollar figure on it. That number makes it easier to decide whether to fix, upgrade or move.
When should I get help?
If you’ve found the pattern and the fix is clear, you may be able to sort it yourself. If the logs are confusing or the outages continue, it’s time for someone who can see the server. Good Australian hosting support can read server logs and resource graphs you can’t access. For ongoing cover, our website maintenance plans include monitoring, updates and repairs. You can learn more about how website uptime monitoring works before you choose.
Frequently asked questions
Why does my website go down for a few minutes and then come back?
Usually because it briefly hits a limit, such as CPU, memory or simultaneous connections, then recovers when the load drops. Traffic spikes, bot attacks and heavy scheduled jobs are the usual triggers. Matching the outage times to these events normally reveals the cause.
What does “508 Resource Limit Is Reached” mean?
It’s an error shown by many shared hosts when your account has too many requests running at once. Caching, blocking bad bots and fixing slow plugins all reduce the load. If the site is genuinely busy, it needs a plan with more resources.
Can a plugin make my website crash?
Yes. A plugin with slow database queries or a memory leak can overload the server, especially under traffic. If outages began after an update or a new plugin, check the error log, then roll it back or replace it.
Is my host responsible for my website going down?
Sometimes. Server faults, network problems and overloaded shared servers are the host’s job. Plugins, themes and a site that has outgrown its plan are usually yours. Ask your host for resource graphs and incident history to tell which.
How do I know when my site goes down?
Use an uptime monitor. It checks your site every minute or so and alerts you by email or text when it fails. It also keeps a history of every outage, which is exactly what you need to find a pattern.
Sources
- 503 Service Unavailable, MDN Web Docs
- Troubleshooting CloudLinux, LiteSpeed Technologies
- 2024 Annual WordPress Security Report, Wordfence, April 2025
- How HTTP status codes and network errors affect Google Search, Google Search Central
- Cron, WordPress Developer Resources