A 500 Error Another Agency Could Not Fix

The problem

• The site returned a 500 error with nothing on the page to indicate why.
• The client’s existing SEO agency had already attempted a fix and could not resolve it.
• The site was only days old, so "it used to work" gave no useful history to compare against.
• A 500 is a symptom, not a diagnosis: it can originate in PHP, the server, or the database.

How we found the cause

• Checked the hosting layer first and found the database, not PHP, was the failure point.
• Went through the database table by table rather than guessing: the fault was a single table consuming almost the entire allowance.
• Identified the plugin writing to it: logging was enabled with no rotation and no cap, on a brand-new site.

The fix

• Took a full database backup before changing anything, then cleared the log data.
• Reconfigured the plugin so its logging could not fill the database again: a clean-up alone would only have delayed a repeat.
• Handed the backup and a written explanation to the client.

The result

• Database reduced from 2,156 MB to roughly 200 MB, back well inside the 2 GB limit.
• Site restored to normal operation.
• Root cause removed rather than the symptom cleared, so the failure could not recur.
• Client received the database backup and the diagnosis in writing.

What this means for you

A 500 error tells you that something failed on the server, not what. It can come from PHP, the web server or the database, and a fix aimed at the wrong layer will not hold, which is where the earlier attempt on this site ended up. Working down the layers in order is what found the cause here.

Hosting plans cap the size of your database, and a plugin with logging switched on and nothing limiting it can fill that allowance quietly, even on a site that is only days old. If you run plugins that keep activity, security or email logs, it is worth knowing where those logs are stored and whether anything stops them growing.

Clearing the data fixed the symptom; changing the plugin’s logging settings is what stopped it coming back. Taking a backup before deleting anything meant the clean-up could have been reversed if it had gone wrong.