PHP New Added 9 September 2026

PHP-FPM: server reached pm.max_children setting, consider raising it

Every worker in the pool is busy, so new requests sit in the listen backlog until Nginx gives up with a 502 or 504. Raising max_children only helps if there is memory to back it: the real cause is usually workers blocked on a slow query or an outbound HTTP call that has no timeout.

Quick fix

Read the commands before running them. Anything that restarts a service, deletes data or changes permissions should be tried on a non-production system first.

Quick fix

# How many workers, and what are they doing right now?
grep -E "pm\.|max_children" /etc/php/8.3/fpm/pool.d/www.conf
curl -s "http://127.0.0.1/status?full"   # needs pm.status_path

# Size the pool from measured memory use rather than a guess
ps -o rss= -C php-fpm | awk '{s+=$1} END {print s/NR/1024 " MB per worker"}'
# max_children is roughly usable RAM divided by that number

# Cap how long one request may hold a worker
request_terminate_timeout = 30s
; php.ini
max_execution_time = 30

# Then give the slow dependency a timeout of its own
curl_setopt($ch, CURLOPT_TIMEOUT, 5);

How to diagnose PHP errors

PHP errors are dominated by per-request resource limits (memory_limit, max_execution_time) and by autoloading. Class-not-found errors are almost never missing code. They are a namespace that does not match the PSR-4 mapping, or a stale Composer autoload map. Laravel adds a caching layer that produces genuinely confusing errors when configuration is cached with the wrong environment.

If the quick fix above does not resolve it, work through these steps. They apply to this whole class of error, not just to this one message, which is usually what saves the time.

Read the real limits from the running process, not from a config file: php -i | grep -E 'memory_limit|max_execution_time'. CLI and FPM have separate configurations.
Regenerate the autoload map with composer dump-autoload -o before investigating a class-not-found error any further.
For Laravel, clear cached config and routes with php artisan optimize:clear. A config cache built in the wrong environment causes errors that survive every code change.
Enable full error display in development (display_errors=On, error_reporting=E_ALL). The default production settings hide the cause.
For PDO connection refusals, test the same credentials with the database's own CLI client to separate PHP configuration from database access.

Tools worth reaching for

php -i
composer dump-autoload -o
php artisan optimize:clear
Xdebug
tail -f storage/logs/laravel.log

Authoritative references

Primary documentation for this error, worth reading before applying any fix in production.

php.net

Related PHP errors

See all 11 PHP errors →

Browse other categories

Something missing or wrong?

This entry is maintained by hand. If the fix is out of date, incomplete, or you have a better one, email a correction and it will be reviewed.