How to find the performance bottlenecks in my Apache configuration?

Apache performance bottleneck inspection needs to start from four aspects: MPM mode, log analysis, Server-status monitoring and module loading. 1. Check and adjust the MPM mode, and reasonably set parameters such as MaxRequestWorkers based on memory; 2. Position slow requests and high-frequency errors through access and error logs; 3. Enable Server-status page to monitor connection status and CPU usage in real time; 4. Disable unnecessary loading modules to reduce resource overhead. During optimization, the effect should be adjusted item by item and observed to ensure that the configuration matches the actual load requirements.

When configuring Apache, performance bottlenecks are often hidden in the details. If you find that the server is slow to respond, high load or poor concurrency, it may not be a hardware problem, but a problem with the Apache configuration itself. To find these bottlenecks, you have to start from several key points.

1. Check Apache's MPM mode settings

Apache has a variety of multi-threaded/multi-process models (MPM), such as prefork , worker , and event . Different modes use different ways of using resources, and choosing the wrong mode or unreasonable configuration will directly affect performance.

• Prefork is a traditional multi-process model, stable but memory consumption.
• Worker and Event use threads to process requests, which is more suitable for high concurrency scenarios.

You can check which MPM you are currently using through the following command:

 httpd -V | grep MPM

Then check the corresponding configuration file (such as /etc/httpd/conf.modules.d/00-mpm.conf or similar path) and adjust the following parameters:

• StartServers
• MinSpareServers / MaxSpareServers
• MaxRequestWorkers
• MaxConnectionsPerChild

The key is to set the maximum concurrency based on your server resources (especially memory). For example, if each Apache child process takes up 20MB of memory and you have 2GB of available memory, then MaxRequestWorkers should not exceed 100.

2. Analyze access logs and error logs

Logs are the most direct source of clues. If you find that some pages are loading very slowly, it may be that the script is executed for a long time, the database query is slow, or the static resource loading is blocked.

• Check if there are a large number of 4xx or 5xx errors in the access log:

   tail -f /var/log/httpd/access_log

• Check whether there are frequent errors in the error log:

   tail -f /var/log/httpd/error_log

Frequently asked questions include:

• Script timeout (PHP, etc.)
• DNS resolution failed
• File permission issues result in 403
• Rewrite rules in .htaccess are inefficient

It is possible to consider combining mod_log_config to record more detailed request time information, such as %D represents the time (milliseconds) for processing the entire request, so that it is easier to find slow requests.

3. Enable and analyze the Server-status page

Apache comes with a status monitoring page that can view current connection status, request status, CPU usage and other information in real time.

The method to enable is simple: open httpd.conf or related configuration files, find and uncomment:

 <Location /server-status>
    SetHandler server-status
    Require host example.com
</Location>

Then visit http://yourdomain.com/server-status to see the real-time running status.

Focus on:

• How many active connections are there currently
• The time when the request is in the state of "W", "K", "_", etc.
• Is the CPU usage abnormally high?

This page can help you determine whether Apache is really “overwhelmed” or is it just a link that is stuck.

4. Check the module loading status

Apache supports many modules, but loading too many unnecessary modules can increase memory and CPU overhead.

Use the following command to view which modules are currently loaded:

 httpd -M

If you see some unused modules, such as mod_perl , mod_php (if you are using PHP-FPM), mod_ssl (if there is no HTTPS request), you can consider disabling them.

The method to disable is usually to comment out LoadModule line in httpd.conf or mods-enabled directory.

Basically that's it. The core of troubleshooting Apache's performance bottleneck is to understand its operating mechanism and resource usage. Some configurations look similar, but vary greatly under actual load. It is recommended to gradually adjust, change one parameter at a time, observe the effect and continue to optimize.

The above is the detailed content of How to find the performance bottlenecks in my Apache configuration?. For more information, please follow other related articles on the PHP Chinese website!