The infrastructure agent is a lightweight piece of software, designed to minimize its impact on the performance of your hosts. However, the exact load varies depending on your host's workload, particularly on the number of processes running on the host. This is because the agent collects detailed data from each individual process.
As a general guideline, New Relic has collected benchmarks for some common types of hosts:
The agent has very low performance overhead on a classic, single-task host. For example, a server running Apache, Unicorn, or a single Java application.
For this type of classic, single-task host, typical usage is:
CPU: about 0.3%
Virtual memory: about 1 GB
Resident memory: 25 to 35 MB
Storage on disk: about 50 MB
The agent has very low performance overhead on a host running Docker, with exact usage depending on the number of Docker containers your machine hosts, and whether those processes are long- or short-lived.
Amazon Linux 2 EC2 instance with infrastructure agent default settings:
CPU: about 0.1% on ARM vs 0.13% AMD
Virtual memory: about 0.75GB ARM vs 1 GB AMD
Resident memory: 20MB ARM vs 22 MB AMD
We are always improving the performance of the infrastructure agent. If you see unusually high agent performance overhead, get support at support.newrelic.com.
Manage data
To learn how to adjust how much data our infrastructure monitoring ingests and reports, see Manage infrastructure data.
Resource utilization
On Linux systems, infrastructure is installed with default settings for each supported service manager. A memory limit of 1 Gigabyte is enforced. Please consider reviewing and adjusting the default configuration based on your system requirements.