Fixing 502 Bad Gateway and 503 Service Unavailable Errors

Introduction

502 Bad Gateway and 503 Service Unavailable are server-side errors. They usually mean the browser reached something — such as a web server, proxy, CDN or load balancer — but the website application, origin server or upstream service could not respond correctly.

A 502 often points to a broken connection between a gateway/proxy and the origin. A 503 usually means the server or application is temporarily unavailable because of overload, maintenance, rate limiting, resource limits or service failure.

Quick answer

502 and 503 errors

Both 502 and 503 are 5xx server-side errors, but they point to different failure patterns.

The browser may show the same error to users, but the fix depends on where the request fails.

502 vs 503

502 causes

503 causes

Error source

Before fixing, identify which layer generated the error.

• browser/client
• CDN edge
• WAF/security layer
• load balancer
• reverse proxy
• web server
• application server
• PHP-FPM/app runtime
• database
• origin hosting server

Error pages often reveal the layer. For example, a CDN-branded 502 may indicate CDN-to-origin failure, while a hosting-branded 503 may indicate resource limits or maintenance.

Why this matters

These errors should be investigated from the outside in: public URL, CDN/proxy, origin server, application service, database and logs.

How to diagnose

Use Website Status Checker to confirm the exact status code, final URL and whether the error is consistent.

Check in this order:

1. Exact URL — Test the affected page, not only the homepage.
2. Error code — Confirm whether it is 502, 503, 504 or another status.
3. Error source — Check whether the page is from CDN, proxy, hosting server or application.
4. CDN/proxy status — Bypass CDN if possible or check origin directly.
5. Origin server — Confirm the hosting server is online and listening on HTTP/HTTPS.
6. Web server logs — Review Nginx/Apache errors.
7. App runtime — Check PHP-FPM, Node, Python, Ruby or other backend services.
8. Database — Confirm database connectivity and load.
9. Resource usage — Check CPU, RAM, disk, processes, workers and limits.

Common problems

How to fix errors

1. Step 1: Confirm the public error

   Use Website Status Checker and record the exact status, URL and response source.

2. Step 2: Check CDN/proxy layer

   Review CDN origin settings, SSL mode, cache and WAF/firewall rules.

3. Step 3: Check origin availability

   Confirm the hosting server responds directly on HTTP/HTTPS.

4. Step 4: Check web server logs

   Review Nginx/Apache error logs for upstream, proxy, timeout or permission errors.

5. Step 5: Check application runtime

   Restart and inspect PHP-FPM, Node, Python, Ruby or other backend services.

6. Step 6: Check database

   Confirm database service is running and not overloaded.

7. Step 7: Check resources and limits

   Review CPU, RAM, disk, I/O, workers, entry processes and account limits.

8. Step 8: Disable bad plugin/deployment if needed

   If the error started after an update, rollback or disable the recent change.

9. Step 9: Monitor after recovery

   Watch logs, response time and resource usage to confirm the error does not return.

Example 1: 502 from CDN

Public check:
https://example.com → 502 Bad Gateway

Origin check:
https://origin.example.com → timeout

Likely cause:
CDN cannot reach origin server.

Fix:
Check origin server status, firewall, CDN origin IP and SSL mode.

Example 2: 503 on shared hosting

Public check:
https://example.com → 503 Service Unavailable

Hosting panel:
Entry processes limit reached.
CPU usage high.

Likely cause:
Account resource limits.

Fix:
Reduce plugin load, cache pages, review bots, optimize PHP/database or upgrade plan.

Example 3: Nginx 502

Nginx error:
connect() to unix:/run/php/php8.2-fpm.sock failed

Likely cause:
PHP-FPM stopped or socket path wrong.

Fix:
Restart PHP-FPM and verify Nginx fastcgi_pass configuration.

Check public status:
curl -I https://example.com

Follow redirects:
curl -IL https://example.com

Check web server status:
systemctl status nginx
systemctl status apache2

Check PHP-FPM:
systemctl status php-fpm
systemctl status php8.2-fpm

Check recent Nginx errors:
tail -100 /var/log/nginx/error.log

Check recent Apache errors:
tail -100 /var/log/apache2/error.log

Check disk:
df -h

Check memory:
free -m

Check load:
uptime

Check processes:
top

WordPress checks

For WordPress sites, 502 and 503 errors often happen after plugin updates, theme changes, heavy traffic or PHP resource exhaustion.

• recently updated plugins
• active theme errors
• PHP fatal errors
• maintenance mode file
• wp-config.php changes
• database connection
• object/cache plugin
• security plugin/WAF rules
• PHP memory limit
• PHP-FPM worker limits
• long-running admin-ajax requests
• WooCommerce or cron load

Hosting checks

On shared hosting, 503 errors may appear when resource limits are reached.

• CPU usage
• physical memory
• virtual memory
• entry processes
• I/O usage
• IOPS
• number of processes
• PHP errors
• Apache/Nginx logs
• ModSecurity blocks
• account suspension or limits

If one account repeatedly hits limits, optimization or a higher hosting plan may be needed.

Server checks

On VPS or dedicated servers, check the full service stack.

• web server running
• PHP-FPM/app service running
• database running
• disk space
• RAM/swap usage
• CPU load
• firewall rules
• fail2ban/WAF blocks
• system logs
• upstream socket/port
• SSL certificate on origin
• proxy configuration

A 502 from Nginx often means the upstream service, socket or port is not responding correctly.

Frequently asked questions