Choose pm mode and child limits from measured memory and traffic patterns.
What It Does and When to Use It
Set parameters in the FPM pool configuration; avoid exposing pool status publicly. The web server handles HTTP, the PHP worker executes scripts, and CLI has separate configuration and lifecycle.
Requirements, Inputs and Expected Behavior
Confirm the runtime, extension, environment variables, and permissions before running the examples in production. Overly large max_children can trigger OOM even when request latency looks good.
Example 1: Basic Operation
Set parameters in the FPM pool configuration; avoid exposing pool status publicly.
pm = ondemand
pm.max_children = 16
pm.process_idle_timeout = 15s
pm.max_requests = 500Example 2: Behavior and Validation
Run the matching status or validation command on the intended host.
php-fpm8.4 -ttPractical Example: Application Use
Correlate capacity, worker crashes and slow requests using logs and process state.
sudo systemctl status php8.4-fpm --no-pager
sudo journalctl -u php8.4-fpm -n 35 --no-pagerImportant Behavior and Edge Cases
The second example highlights: Run the matching status or validation command on the intended host. The practical example adds: Correlate capacity, worker crashes and slow requests using logs and process state.
Comparison with Related Approaches
| Approach | What changes |
|---|---|
| Web server, worker and PHP CLI | The web server handles HTTP, the PHP worker executes scripts, and CLI has separate configuration and lifecycle. |
| Choosing safely | Check the effective configuration in the SAPI serving requests before changing production. |
Common Beginner Mistakes
| Mistake | Correct approach |
|---|---|
| Checking the PHP CLI INI instead of FPM | Inspect phpinfo or configuration through the protected web SAPI. |
| Exposing worker status publicly | Restrict FPM diagnostics to localhost or a monitored private network. |
| Skipping operational constraints | Overly large max_children can trigger OOM even when request latency looks good. |
Key Takeaways
- Choose pm mode and child limits from measured memory and traffic patterns.
- Correlate capacity, worker crashes and slow requests using logs and process state.
- Check the effective configuration in the SAPI serving requests before changing production.