An error starting with 5 carries good news inside it: something answered. The server took the request, tried to build the page and fell over halfway. Had there been nobody on the other end, the browser would never have shown you a number at all.
That is why these are worth reading properly: the number tells you which side to start looking at. Codes 500, 403, 404 and 508 have their own guide, in website errors explained. This one covers the whole family, and above all the 503, which is the one missing there.
The 5xx family, two lines each
| Code |
What it actually means |
| 500 |
The server tried to run the site and blew up. Almost always .htaccess, permissions, or the application’s own code. It has its own section in website errors explained. |
| 502 |
Whoever was in front got a broken answer from whoever was behind. If you have a CDN or any middleman in the path, start there. |
| 503 |
The service that should have answered was not available. The web server is alive, obviously, since it answered you: what failed to start was the process that runs your site. |
| 504 |
Whoever was behind took longer than allowed. Usually a heavy database query, or a process that got stuck. |
| 508 |
The account hit the plan’s resource ceiling. Not a fault: a brake. Also covered in website errors explained. |
Why 503 is the common one on shared hosting
On a shared plan your site does not run on its own: it runs inside a space with a ceiling. When the site asks for more processes, more memory or more time than that space allows, the next request finds nobody free to serve it. The web server still answers, because it is up, and what it answers is 503.
Hence the symptom that confuses everyone: the site comes and goes. Reload and it opens. Reload again and it errors. A broken site never works. A site hitting its ceiling works when there is room and fails when there is not.
The three origins, and how to tell them apart
| Where it comes from |
How to recognise it |
What to do |
| The machine or the service |
Everything fails at once, yours and other sites, and you changed nothing. |
That one is ours. Open a ticket with the exact time. What counts as being down is in uptime. |
| The application |
It started right after an update, a new plugin or a change of yours. It fails every time, and only on your site. |
The cause is already written down: read the PHP error log. |
| The account ceiling |
It comes and goes. Worse at busy hours. It started when the site grew, or when someone began hammering it. |
Check your usage in cPanel, then what actually makes a difference. |
Where to start, in order
| 1 |
Confirm it is not only you. Open it in a private window and on your phone over mobile data. If it is fine there, the error is not on the server: it is stored somewhere along the way. See clearing the cache.
|
|
| 2 |
Look at who signed the error page. With a CDN or any middleman in front, the error may be theirs and not ours. The page usually says who produced it. See Cloudflare and the cache that fools you.
|
|
| 3 |
Check your resource usage in cPanel. It is the fastest answer to “did I hit the ceiling?”. Look for the resource usage section.
|
|
| 5 |
Undo the last thing you changed. If it started after an update, go back first and investigate afterwards. Recovering is faster than understanding.
|
|
|
A 5xx is not the same as “the site will not open”. If the browser shows no number at all and says it could not connect, then nothing answered and the problem is a different one. That case is in when the error is not the server’s, and the full checklist is in your site is down.
|
If you write to us, three things cut the answer from hours to minutes: the code, the exact address of the page that failed, and the time you saw it fail. With those we go into the server logs.
RECOMMENDED PRODUCT Web hosting with cPanel Domain and SSL included, daily backups and the panel you already know. from $10.00/mo See plans |