Before Installation

Before installing Fluentd, make sure that your environment is properly set up to avoid any inconsistencies at a later stage.

Follow these recommendations:

  • Set Up NTP

  • Increase the Maximum Number of File Descriptors

  • Optimize the Network Kernel Parameters

Set Up NTP

It is highly recommended that you set up an NTP daemon (e.g. chrony, ntpd, etc.) on the node to have an accurate current timestamp. This is crucial for all the production-grade logging services.

For Amazon Web Services users, we recommend using AWS-hosted NTP server.

Increase the Maximum Number of File Descriptors

Increase the maximum number of file descriptors. You can check the existing configuration using the ulimit -n command:

$ ulimit -n
65535

If your console shows 1024, it is insufficient. Please add the following lines to your /etc/security/limits.conf file and reboot your machine:

root soft nofile 65536
root hard nofile 65536
* soft nofile 65536
* hard nofile 65536

If you are running fluentd under systemd, the option LimitNOFILE=65536 can also be used. And, if you are using the td-agent package, this value is set up by default.

Optimize the Network Kernel Parameters

For high load environments with many Fluentd instances, add the following configuration to your /etc/sysctl.conf file:

net.core.somaxconn = 1024
net.core.netdev_max_backlog = 5000
net.core.rmem_max = 16777216
net.core.wmem_max = 16777216
net.ipv4.tcp_wmem = 4096 12582912 16777216
net.ipv4.tcp_rmem = 4096 12582912 16777216
net.ipv4.tcp_max_syn_backlog = 8096
net.ipv4.tcp_slow_start_after_idle = 0
net.ipv4.tcp_tw_reuse = 1
net.ipv4.ip_local_port_range = 10240 65535

Use sysctl -p command or reboot your node for the changes to take effect.

These kernel options were originally taken from the presentation "How Netflix Tunes EC2 Instances for Performance" by Brendan Gregg, Senior Performance Architect at AWS re:Invent 2017.

If this article is incorrect or outdated, or omits critical information, please let us know. Fluentd is an open-source project under Cloud Native Computing Foundation (CNCF). All components are available under the Apache 2 License.