If your WordPress backend is not working, first determine what “not working” means. A completely inaccessible /wp-admin/ points to a different problem than a dashboard that opens but cannot save posts, load menus, upload media, or respond to buttons.
In this guide
Start with the least-invasive checks: test a private browser window, confirm whether the public site works, and identify whether the failure affects one account, one browser, one admin screen, or every administrator. Do not immediately edit wp-config.php, replace .htaccess, or repair the database.
This walkthrough helps you isolate the failing layer—browser, WordPress plugin or theme, REST/AJAX request, user permissions, PHP, or hosting—then verify the fix safely.
What “backend not working” usually looks like
Record the exact symptom before changing anything. Useful examples include:
- The login page refreshes without logging you in.
- The dashboard opens, but screens remain stuck loading.
- The block editor cannot save or displays a publishing error.
- Buttons, menus, media uploads, or settings controls do nothing.
- You see a blank page, a critical-error message, or a server error.
- Only one administrator or one browser has the problem.
- The public website works, but authenticated WordPress requests fail.
Note the affected URL, time of failure, user account, browser, and any visible error. A screenshot is useful, but also copy the text of the error because a screenshot may omit an HTTP status or browser-console detail.
Run these safe checks first
1. Confirm the scope of the failure
Open the public homepage in a separate tab. Then test the affected backend screen in a private or incognito window. If the private window works, stale cookies, cached JavaScript, a browser extension, or a session issue is more likely than a server failure.
Also test a second browser or device. Disable extensions temporarily, especially ad blockers, script blockers, password managers, and security extensions. Do not delete all browser data before recording which test changed the behavior.
2. Check whether the problem affects every administrator
If another administrator can use the same screen, the issue may involve your browser session, user capability, account-specific metadata, or an account restriction. If all administrators see the same failure, investigate the site rather than changing your personal browser.
Do not create a new administrator account casually on a live site. If you need an account test, use an approved temporary account, protect it with a strong unique password, and remove it after testing.
3. Avoid repeated updates and retries
If saving or updating fails, do not click the button repeatedly. Repeated requests can create duplicate content, partial settings changes, or multiple orders in an ecommerce system. Check whether the change actually completed in another tab or on the public site first.
Separate a screen problem from a request problem
Many backend controls depend on JavaScript requests, the WordPress REST API, or AJAX. The page can appear normal while the request that saves content or loads data is failing.
Use the browser’s developer tools
- Open the affected backend screen.
- Open your browser developer tools, usually with
F12or the browser’s “Inspect” command. - Choose the Console and Network panels.
- Reload the page, then reproduce the failure once.
- Look for failed requests and record their status, such as
401,403,404,500, or a request that remains pending.
A 403 often suggests a firewall, security rule, or permissions problem. A 401 can indicate authentication or nonce failure. A 500 points toward PHP or server-side execution. A JavaScript error may identify the plugin or script that stopped the interface, but do not assume the first red error is the root cause.
If the REST API is blocked, the editor and other administrative features may fail even though the homepage loads. Compare the symptom with safe ways to diagnose a blocked WordPress REST API.
Test for a plugin or theme conflict

A plugin update, newly added integration, optimization setting, or theme change can break only backend requests. The safest confirmation is a temporary conflict test, not permanent deletion.
If you can access the Plugins screen
- Make a current backup or confirm that your host has a restorable backup.
- Deactivate recently updated or recently installed plugins first.
- Test the exact backend action that failed.
- If the problem remains, deactivate other plugins as a group only if the site can tolerate the temporary change.
- Reactivate plugins one at a time, testing after each activation.
When one plugin restores the failure, leave it inactive while checking its support notes, logs, configuration, and compatibility with the current WordPress and PHP versions. Do not assume the last plugin updated is responsible; another plugin may have changed a shared library or security rule.
If you cannot use the Plugins screen
Use your hosting file manager or SFTP only if you are comfortable identifying the correct WordPress installation. After making a backup, rename the specific plugin directory under wp-content/plugins so WordPress treats it as inactive. Rename only the suspected directory, not the entire wp-content folder.
This is a production file change. Confirm the path with your host, preserve the original directory name so it can be restored, and avoid editing or deleting plugin files. If the backend returns after the rename, restore the directory name and use the dashboard or a controlled maintenance window to isolate the plugin properly.
If all plugins are ruled out, a theme can be involved through admin scripts, editor support, or custom hooks. Test with a default WordPress theme only when one is already installed and you understand the temporary effect on the public site. Do not switch a live ecommerce theme without checking its layout and checkout behavior.
Check WordPress, PHP, and server evidence
Review the recovery email and critical-error details
When WordPress detects a fatal PHP error, it may send the site administrator a recovery email containing the affected plugin or theme and a recovery link. Check the site mailbox, spam folder, and hosting mail logs if no message arrives.
The error details may name a file and line number. That identifies where execution stopped, but not always which component caused it. Record the details before changing files. The official WordPress Documentation is useful for confirming current recovery and troubleshooting procedures.
For a broader explanation of this particular message, see how to fix “There Has Been a Critical Error on This Website”.
Use debug logging carefully
If you have a staging site, enable debugging there first. On production, logging can expose paths, queries, usernames, or other details, so use a temporary configuration and protect the log file. Avoid displaying errors to visitors.
WordPress debug settings are commonly controlled in wp-config.php, but exact hosting arrangements vary. Take a backup and ask your host or developer to confirm the correct method before editing that file. After collecting the evidence, disable temporary debugging and remove or secure logs.
Check the hosting error log
Your hosting control panel may provide PHP, web-server, or application logs. Look at the timestamp that matches your failed backend action. Search for memory exhaustion, maximum execution time, permission errors, missing files, PHP exceptions, or upstream connection failures.
If the log shows a PHP version incompatibility, compare the site’s required versions with the versions supported by its active plugins and theme. Change PHP versions only after confirming a backup, staging option, and host-supported rollback path. A PHP change can affect the entire site.
Check permissions, security rules, and caching
Permissions and ownership
If the backend cannot create uploads, update files, or save settings, incorrect file ownership or permissions may be involved. A permissions change after migration, a server move, or manual FTP upload is a common clue.
Do not recursively apply guessed permission commands to the entire site. Incorrect ownership or permissions can expose configuration files or prevent WordPress from working. Ask the host to compare the affected paths with the server’s normal WordPress ownership and permission policy. For background, see the practical guide to WordPress .htaccess before changing server rules.

Security plugins, WAFs, and hosting firewalls
A security plugin, CDN, web application firewall, login protection rule, or host firewall may block REST, AJAX, heartbeat, uploads, or authenticated requests. Look for a blocked-request event at the exact time of the failure.
Do not permanently disable security controls as a first response. Use a temporary, narrow allow rule for the affected path or account only when the provider documents the process, then remove it after testing. If the site is also redirecting visitors or showing unfamiliar files, stop making random changes and follow a malware response plan such as safe WordPress malware-removal steps.
Clear the right cache
Clear the browser cache or use a private window first. Then, if appropriate, purge the WordPress page or object cache, CDN cache, and host cache one layer at a time. Backend pages should not be served from a public page cache; ask the cache provider to verify that /wp-admin/, /wp-login.php, and authenticated cookies are excluded.
Do not purge every cache repeatedly while troubleshooting. It can hide the timing of the fix and create unnecessary load.
Use this decision guide
- Only one browser fails: test extensions, private browsing, cookies, and browser console errors.
- Only one user fails: test capabilities, account metadata, session cookies, and security restrictions.
- The editor cannot save: inspect REST or AJAX requests, permalinks, firewall events, and nonce errors.
- Every backend page fails: check PHP fatal errors, server logs, permissions, disk space, and recent deployments.
- Only uploads fail: check media directory permissions, PHP upload limits, available storage, and image processing errors.
- The public site also fails: prioritize hosting, PHP, database, DNS, or malware investigation instead of browser troubleshooting.
Verify the fix without creating a new problem
- Repeat the original failing action, not just a dashboard reload.
- Test in your normal browser and a private window.
- Check both an administrator account and a lower-privilege account when permissions are relevant.
- Confirm that the public site, forms, media, ecommerce functions, and scheduled tasks still work.
- Review logs after the test to ensure the same error is no longer being generated.
- Restore any temporary plugin, file, firewall, or debug change that was not needed.
If the problem returns immediately after reactivating a plugin or switching back to a theme, keep the confirmed cause isolated and update or replace it through a staging or maintenance process. Do not repeatedly reinstall WordPress core; that will not fix a plugin conflict, blocked request, or server configuration problem.
When to ask for professional help
Get host or WordPress specialist assistance when the site is losing data, the database is involved, PHP errors continue after isolation, the server cannot write files, or you need to edit production configuration. A specialist can correlate web-server, PHP, firewall, and WordPress logs without guessing.
For a live site that is blocking customers or administrators, WordPress site repair can help investigate the failure while preserving a recovery path.
Questions that often reveal the cause
Why does the backend fail while the homepage works?
The homepage may be served from cache and use few authenticated requests. The backend relies on PHP, logged-in cookies, REST, AJAX, database writes, and security rules. Inspect the failed backend request and its server log rather than assuming the site is healthy because the homepage loads.
Why does the problem disappear in a private window?
That strongly suggests a browser-specific issue such as stale cookies, cached scripts, an extension, or an expired session. Clear only the affected site’s data, disable extensions one at a time, and sign in again. If the issue returns in every browser, investigate the site or host.
Can I fix a broken backend by renaming every plugin folder?
It can be a useful emergency isolation test, but it changes the live site and may disable security, caching, ecommerce, or other essential functions. Back up first, document the change, rename only the intended directory where possible, and restore services promptly after testing.
Should I repair the database when WordPress buttons do nothing?
Not usually. Unresponsive buttons more often involve JavaScript, REST/AJAX, a firewall, or a plugin conflict. Consider database repair only when logs or database symptoms support it, and never run a repair or restore without a verified backup and a clear rollback plan.



