How to Tune MaxRequestWorkers in Apache?

Tuning the MaxRequestWorkers directive (formerly known as MaxClients) is critical for optimizing Apache HTTP Server performance, preventing server crashes, and managing memory usage under high traffic. This guide covers how to calculate the optimal value for your server based on available RAM, configure the directive within your Multi-Processing Module (MPM), and apply the changes safely.

Understanding MaxRequestWorkers

The MaxRequestWorkers directive sets the limit on the number of simultaneous client requests that Apache will serve. Any connection attempts beyond this limit are queued up to the number defined by the ListenBacklog directive.

If the value is too low: Your server will underutilize its resources, and users may experience slow loading times or connection timeouts because they are stuck in the queue.
If the value is too high: Apache will attempt to handle more concurrent requests than the server’s physical memory (RAM) can support, leading to heavy use of swap space, severe performance degradation, or an Out-Of-Memory (OOM) crash.

Step 1: Calculate Your Server’s Capacity

To find the right value, you need to know your total available RAM and the average memory size of a single Apache process.

Find the average RAM used by an Apache process: You can run the following command in your terminal to see the average resident set size (RSS) of your Apache processes in megabytes:

ps -ylC apache2 --sort:rss | awk '{sum+=$8; ++n} END {print "Avg: "sum/n/1024" MB"}'

(Note: Replace apache2 with httpd depending on your Linux distribution, such as CentOS/RHEL). 2. Determine available RAM for Apache: Deduct the memory required for your operating system and other services (like MySQL, PHP-FPM, or Redis) from your total RAM. If your server is a dedicated web server, you can safely allocate roughly 80% of the total RAM to Apache. 3. Use the Tuning Formula: Divide the available RAM by the average Apache process size: \[MaxRequestWorkers = \frac{\text{Available RAM for Apache}}{\text{Average Apache Process Size}}\]

For example, if you have 8 GB of total RAM, you might reserve 2 GB for the OS and other services, leaving 6 GB (6144 MB) for Apache. If your average Apache process size is 50 MB: \[6144 \text{ MB} / 50 \text{ MB} = 122.88\]

In this scenario, you would round down and set your MaxRequestWorkers to 120.

Step 2: Configure the Directive in Apache

Apache handles concurrency via Multi-Processing Modules (MPMs), usually event, worker, or prefork. You must place the directive inside the configuration block for the specific MPM your server is running.

Open your Apache configuration file (commonly found at /etc/apache2/mods-enabled/mpm_event.conf, /etc/apache2/apache2.conf, or /etc/httpd/conf.modules.d/00-mpm.conf).
Locate or add the directive inside your active MPM block.

For the modern Event or Worker MPMs, you must also ensure that ServerLimit is configured properly, as MaxRequestWorkers cannot exceed ServerLimit divided by ThreadsPerChild.

<IfModule mpm_event_module>
    ServerLimit           16
    ThreadsPerChild       25
    MaxRequestWorkers     400
    MaxConnectionsPerChild 5000
</IfModule>

For the legacy Prefork MPM (where one process handles one connection), the configuration is simpler:

<IfModule mpm_prefork_module>
    ServerLimit           120
    MaxRequestWorkers     120
    MaxConnectionsPerChild 3000
</IfModule>

Step 3: Test and Restart Apache

Before applying the new limits, always test your Apache configuration syntax to ensure there are no typos or structural errors.

apachectl configtest

httpd -t

If the output returns Syntax OK, restart the Apache service to apply the changes:

sudo systemctl restart apache2

sudo systemctl restart httpd