Performance optimisations

This document gives guidance on the optimisation of your setup.

The list starts with the changes that have the largest effect.

RRDCached

We recommend RRDCached. It reduces the IO load.

MySQL Optimisation

After 24 hours of MySQL operation, run MySQL Tuner. This tool recommends changes for your own setup.

We recommend this setting in my.cnf, under a [mysqld] group:

innodb_flush_log_at_trx_commit = 0

You can also set this option to 2. A crash of MySQL or of the server can then lose up to 1 second of MySQL data. In exchange, the IO use decreases a lot.

Polling modules

Open gear > pollers > performance. The graph of the poller module time shows which modules use the poller time. The same data is available per device under device > graphs > poller.

Disable the polling modules and the discovery modules that you do not need. To disable a module globally, use lnms config:set poller_module.<module>.

To disable OSPF polling, use this command:

lnms config:set poller_modules.ospf false

You can disable a module globally, then enable it for one device. You can also do the opposite. For a list of the modules, read Poller modules.

SNMP Max Repeaters

LibreNMS supports SNMP max repeaters. This option helps on devices with many ports or many BGP sessions, where LibreNMS uses snmpwalk or snmpbulkwalk. Enable the option for one device under edit device -> snmp -> Max repeaters.

You can also set the option globally:

lnms config:set snmp.max_repeaters X

To find the best value, measure the time of an snmpwalk over IF-MIB. Run the command below with different values in place of -REPEATERS-, from 10 to about 50. Also set the correct SNMP version, hostname, and community string:

time snmpbulkwalk -v2c -cpublic HOSTNAME -Cr-REPEATERS- -M /opt/librenms/mibs -m IF-MIB IfEntry

SNMP Max OIDs

For sensor polling, LibreNMS does bulk snmpget requests. This makes the polling faster. The default value is ten. You can change the value for one device under edit device -> snmp -> Max OIDs.

You can also set the option globally:

lnms config:set snmp.max_oid X

fping tuning

You can change the default fping options globally or for one device. These are the defaults:

lnms config:set fping_options.timeout 500
lnms config:set fping_options.count 3
lnms config:set fping_options.interval 500

If your devices answer slowly, increase the timeout value. You can also increase the interval value. If your network is stable, you can improve the poller performance. Decrease the count value to 1, the timeout value to 200 or 300, or both:

lnms config:set fping_options.timeout 300
lnms config:set fping_options.count 1
lnms config:set fping_options.interval 300

LibreNMS then sends no delay of 0.5 seconds between the ICMP packets. By default, LibreNMS sends 3 packets in total. With only 1 ICMP packet, the response arrives faster. With the defaults, a response takes at least 1 second, even for a fast device.

Optimise poller-wrapper

The default for poller-wrapper.py is 16 threads. This value is not always optimal. A common rule is 2 threads for each core. Increase and decrease the number until you find the optimal value.

*/5  *    * * *   librenms    /opt/librenms/cronic /opt/librenms/poller-wrapper.py 16

If you use the Dispatcher Service, you can set the number of threads in the web interface. For more information, read Dispatcher Service.

Recursive DNS

If your install uses hostnames for many devices, install a local recursive DNS instance on the LibreNMS server. You can use pdns-recursor. Then configure /etc/resolv.conf to send queries to 127.0.0.1.

Per port polling

By default, the ports polling module walks ifXEntry and some items from ifEntry. It does this walk for every port, at every port status. LibreNMS therefore collects data for a deleted port and for a disabled port. The walks are fast, so this behaviour is usually acceptable. It is not optimal on a device with many ports where a large percentage is deleted or disabled. For such a device, enable "selected port polling" under edit device -> misc. You can also enable it globally, but we do not recommend this:

lnms config:set polling.selected_ports true

We do not recommend the global setting, because it increases the CPU use of your poller. You can also set the option for one OS:

lnms config:set os.ios.polling.selected_ports true

Run ./scripts/collect-port-polling.php as the librenms user. The script polls your devices with full polling and with selective polling. It then shows a table with the difference. It can also enable or disable selected ports polling for the devices with a benefit. The script does not repeat this analysis. It updates the setting only at each run. These options are available:

-h <device id> | <device hostname wildcard>  Poll single device or wildcard hostname
-e <percentage>                              Enable/disable selected ports polling for devices which would benefit <percentage> from a change

To set selected port polling on the devices with a change of 10% or more, run ./scripts/collect-port-polling.php -e 10. The script uses a second condition. The change in the polling time must also be more than one second.

Web interface

HTTP/2

If you use HTTPS, enable HTTP/2 support in your web server:

For Nginx 1.9.5 and later, change listen 443 ssl; to listen 443 ssl http2; in the virtual host configuration.

For Apache 2.4.17 and later, set Protocols h2 http/1.1 in the virtual host configuration.

PHP-opcache

A correct php-opcache configuration gives a large gain in performance.

Note: memory based caching with the PHP command line increases the memory use and makes the system slower. File based caching is slower than memory based caching, and stale cache problems are more common.

Some distributions have separate configurations for the command line, mod_php, and php-fpm. Use these to set the optimal configuration.

For web servers using mod_php and php-fpm

Update the opcache.ini file of your web PHP. The possible locations are /etc/php/8.3/fpm/conf.d/opcache.ini, /etc/php.d/opcache.ini, and /etc/php/conf.d/opcache.ini.

zend_extension=opcache
opcache.enable=1
opcache.memory_consumption=256

If you have a caching problem, restart httpd or php-fpm. This clears the opcache.

For pollers

First create a cache directory. The librenms user must have write access to it: sudo mkdir -p /tmp/cache && sudo chmod 775 /tmp/cache && sudo chown -R librenms /tmp/cache

Update your opcache.ini file. The possible locations are /etc/php/8.3/cli/conf.d/opcache.ini, /etc/php.d/opcache.ini, and /etc/php/conf.d/opcache.ini.

zend_extension=opcache.so
opcache.enable=1
opcache.enable_cli=1
opcache.file_cache="/tmp/cache/"
opcache.file_cache_only=0
opcache.file_cache_consistency_checks=1
opcache.memory_consumption=256

If you have a caching problem, clear the file based opcache with rm -rf /tmp/cache.

On Debian 12, PHP 8.2.7 causes segmentation faults when opcache uses the file cache. The cause is this problem: https://github.com/php/php-src/issues/10914 To correct it, use the sury packages or disable the file cache.