Unlock Effortless Kubernetes Monitoring with eAPM

Part 1: Introduction

Managing Kubernetes workloads often involves complex manual instrumentation and constant lifecycle management of multiple agents, significantly impacting engineering productivity. Traditional monitoring solutions can be cumbersome, requiring extensive code changes and manual updates to ensure comprehensive visibility.

This can now be tackled with New Relic’s eAPM—an innovative monitoring solution leveraging extended Berkeley Packet Filter (eBPF) technology to detect and monitor Kubernetes-based workloads automatically. eAPM eliminates manual instrumentation with zero-code integration, delivering comprehensive, real-time performance insights directly into the New Relic One platform, significantly enhancing your observability capabilities.

Part 2: Supporting Details

No-Code Instrumentation: Automatic Discovery and Instrumentation

Traditionally, achieving full-stack observability required manual instrumentation and frequent agent updates. eAPM addresses this by automatically detecting and instrumenting all services and applications running in your Kubernetes clusters without any code changes. This significantly streamlines workflow, reduces deployment complexity, and eliminates maintenance overhead.

How we do it:
With a simple Helm command or guided installation, eAPM deploys an efficient eBPF agent directly on each Kubernetes node, automatically capturing and sending telemetry data to New Relic’s APM interface, regardless of the application’s programming language.

Real-Time Debugging: Intelligent Span Sampling

Identifying the root cause of performance issues in complex Kubernetes environments can be a time-consuming and challenging task. eAPM utilizes intelligent span sampling to identify and highlight critical performance issues within application spans automatically. This feature simplifies debugging by quickly surfacing the essential information required to resolve performance bottlenecks.

Practical example:
If a critical microservice suddenly experiences latency, eAPM quickly highlights problematic spans, enabling the rapid identification and resolution of the issue, thereby minimizing downtime and improving the user experience.

AI-Powered Insights: Comprehensive Observability

Traditional monitoring solutions often require manual correlation across multiple data points, which slows incident response. eAPM provides comprehensive AI-driven insights directly within the New Relic APM UI, automatically correlating golden metrics, transaction details, database performance, and Kubernetes cluster performance.

Benefits:

• Instant visibility into the health and performance of your entire Kubernetes workload.
• Reduced manual effort through automatic correlation of telemetry data.
• Enhanced ability to proactively address potential performance degradation.

Resource-Efficient Operation: Minimal Overhead

Monitoring solutions have traditionally consumed substantial system resources, negatively impacting overall application performance. eAPM operates efficiently at the kernel level using eBPF technology, ensuring low CPU and memory overhead, allowing your applications to run optimally without additional performance penalties.

Impact:
Resource optimization enables your Kubernetes clusters to perform at their best, freeing up valuable resources for critical business workloads rather than dedicating them to monitoring overhead.

Seamless Transition to Full APM: Flexibility and Comprehensive Coverage

Other eBPF-based monitoring solutions often lock users into limited monitoring capabilities, making transitions cumbersome. New Relic’s eAPM allows for seamless upgrading to full APM agents when deeper monitoring insights, such as distributed tracing, are required. This seamless transition ensures continuous observability, without disruption or the need to recreate dashboards and alerts.

Result:
Smooth operational continuity, providing the flexibility to scale your observability solutions according to evolving business needs.

Part 3: Getting Started

Deploying eAPM to your Kubernetes cluster is straightforward with New Relic’s guided installation. To deploy eAPM:

1.     Ensure your Kubernetes nodes are running a Linux kernel version 4.14 or higher.

2.     Use the guided installation or run a Helm command to install the eBPF agent on each node.

3.     Navigate to the New Relic One platform and navigate to APM and Services [one.newrelic.com> APM > Services] to view your auto-discovered services and performance data.

From there, you can evaluate which services may benefit from full APM instrumentation and transition seamlessly.