Monitoring with Agent Control

Overview

Agent Control is designed to be a local supervisor, while New Relic Control provides the centralized monitoring experience. Agent Control acts as the on-cluster supervisor for your managed agents. Its primary role is to ensure that agents are correctly deployed and running. This guide details how to monitor your managed agents in a Kubernetes environment. For a high-level view of your fleet's health, you can use the New Relic Control UI. For in-depth troubleshooting and validation, you can perform checks directly on your cluster.

On-cluster health checks and troubleshooting

These checks are performed using standard Kubernetes tools and commands.

When troubleshooting an issue, you can inspect the status of Agent Control and its managed agents directly using kubectl and the local status endpoint.

Check pod status and logs

You can use standard Kubernetes commands to verify that the Agent Control supervisor and its sub-agent pods are running as expected.

$
# Check Agent Control pods
$
kubectl get pods -n newrelic-agent-control
$

$
# Check sub-agent pods (e.g., in the 'newrelic' namespace)
$
kubectl get pods -n newrelic
$

$
# Inspect logs for the Agent Control supervisor
$
kubectl logs <agent-control-pod-name> -n newrelic-agent-control

Use the local status endpoint

Agent Control exposes a local status endpoint that provides a health check of both the supervisor and all its managed agents.

1. Forward the port: Open a terminal and run the following command to forward the local port to the agent-control pod.

$
kubectl port-forward <agent-control-pod-name> 51200:51200 -n newrelic-agent-control

2. Request the status: In a new terminal, use curl to query the status endpoint.

$
curl localhost:51200/status

The output is a JSON response that provides a detailed health status, including deployment details, agent versions, and any detected errors.

Centralized monitoring in New Relic Control

The ultimate goal of the Agent Control solution is to provide a comprehensive, centralized monitoring experience within the New Relic UI. This will move you away from needing to perform manual kubectl commands for health checks and give you a unified view of your entire fleet. This experience will be available for all environments supported by Agent Control, including the upcoming host-based support.

In the New Relic Control UI you can manage your fleets and see their status, including the status of each managed agent (e.g., Healthy, Unhealthy, Applying).

This allows you to quickly transition from monitoring the health of a fleet to investigating the specific metrics, logs, and traces collected by an individual agent, providing a seamless workflow from high-level management to deep-dive observability.