New Relic's on-host integrations are a type of infrastructure integration that collect data from core services running on your hosts, such as MySQL, Apache, or Redis, among others.
There are multiple ways to install on-host integrations depending on your setup and needs. Here we present a brief overview of all install methods and when it's more appropriate to use them.
What do you need?
To collect data from on-host services, all you need are two things: the infrastructure agent, and the on-host integrations. All on-host integrations require that our infrastructure agent is installed on the host. Besides collecting system's data, the agent acts as a forwarder for integrations's data, and can forward logs.
The infrastructure agent and its integrations collect data from the system and core services, and is able to forward logs to New Relic. Backend application metrics (APM) are collected by separate APM agents. Notice how each integration and forwarder feed different data types in the New Relic database (NRDB).
Quick start
The quickest way to get started is through our guided install.
Your services may be running on a single physical host, on a VM, or in a container orchestrated by Kubernetes or ECS. Whatever the setup, our on-host integrations can adapt to your environment and send data to New Relic; all you have to do is choose the appropriate install method.
You're running services in orchestrated environments
If you are running services in containerized, orchestrated environments, choose your scenario:
In an admin account, run the install script using an absolute path.
msiexec.exe /qn /i PATH\TO\integration-name.msi
Change C:\Program Files\New Relic\newrelic-infra\integrations.d\integration-name-config.yml.sample to integration-name-config.yml, and edit according to your needs.
Run the following command, where INTEGRATION_FILE_NAME represents the integration's file name. For more information, see the integration's documentation.
Run the following command, where INTEGRATION_FILE_NAME represents the integration's file name. For more information, see the specific on-host integration.
Run the following command, where INTEGRATION_FILE_NAME represents the integration's file name. For more information, see the integration's documentation.
The manual install process is not automated. If you opt for manual install, you must place the different files in the correct folders, and ensure that the agent has all the permissions to execute the integrations.
Unpack the tarball file according to our integration file structure and placement rules, so the agent is able to find the definitions, configurations, and executables of the integration.
Place the binary that contains the definition file inside newrelic-integrations or custom-integrations in the agent directory.
For integrations that require our nrjmx tool, follow these additional instructions:
Use of the New Relic JMX tool
Some integrations (such as JMX, Cassandra, and Kafka) require the nrjmx tool. If your integration needs this, download it from our repository and unpack it.
Important
nrjmx requires Java 8 or higher.
For JMX integration version 2.3.3 or higher and Cassandra integration version 2.3.0 or higher, the nrjmx tool is included as a dependency. For this reason, when using a package manager, the nrjmx tool doesn't have to be installed manually.
If you have nrjmx already installed and install nri-jmx, our JMX tool keeps the already-installed version. If you don't have nrjmx already installed, it gets the latest nrjmx release.
By default, the nrjmx location is /usr/bin/nrjmx/*. To install in a different location, set the new path in the NR_JMX_TOOL environment variable.
Install on-host integrations on Kubernetes
Install on-host integrations on Amazon ECS (EC2 launch type)