Kubernetes summary page

The Kubernetes APM summary page provides general information about the selected application, including the age of the current deployment, the number of pods available, CPU and memory usage, pod logs, and metric comparison.

Monitor the deployment of your Kubernetes cluster

The Kubernetes page in APM offers the following information about your application:

Deployment summary

You can see your cluster name, deployment, and namespace that it's running in. If you click the deployment, you'll get details about it in a slide-out panel. In the slide-out panel, you can see a mini overview containing relevant information such as a count of logs and percentage of error logs in the specified time range.

In the Instances drop-down, you can select a single pod instance of your service.

If you have a multi-cluster app (an app that spans multiple clusters), the Kubernetes page shows you a summary of all the clusters. To isolate a specific cluster, select it from the Cluster drop-down list. To return to a view of all clusters, click All clusters.

Current deployment

You can check how many pods are available or desired, the pods that are missing, and the age of the kubernetes deployment, which is the date when the deployment was first created.

Performance summary for the deployment

You can see the CPU and memory utilization, as well as the number of pod restarts.

Pods and containers

You can check pod and their related containers details. If you click the pod or container name, you'll get some details about it in a slide-out panel.

In addition, with the Logs button, you can see the logs in context to that pod.

Activity stream

You can see the stream of Kubernetes events related to that deployment and its pods.

Metrics comparison

These widgets let you compare a static list of metrics related to the pods-containers in the deployment, as well as other metrics such as CPU, network, storage.

The following metrics are available:

• CPU usage (cores)
• CPU throttling (%)
• Memory usage (bytes)
• Network received (KBps)
• Network transmitted (KBps)
• Network errors (per second)
• Storage usage (bytes)
• Storage utilization (%)
• Container restart count
• Service throughtput (requests per minute)
• Service response time (ms)
• Service error rate (%)

Link your applications to Kubernetes

You can surface Kubernetes metadata and link it to your APM agents, whether you use our proprietary agents or OpenTelemetry. APM uses the Kubernetes metadata to display the specific Kubernetes deployment information related to that APM application.

New Relic proprietary instrumentation

When you install New Relic's Kubernetes integration, New Relic's metadata injection comes into play automatically via the nri-metadata-injection parameter. The MutatingAdmissionWebhook, set up by the Kubernetes instrumentation, takes care of tagging your pods with the necessary environment variables. From there, the data captured in these variables gets carried over to your APM metrics and entities, enriching them with valuable context.

For more information, see how to link your applications to Kubernetes.

The following attributes are required in the APM service entity to display the Kubernetes summary page:

• k8s.clusterName
• k8s.namespaceName
• k8s.deploymentName

OpenTelemetry instrumentation

The OpenTelemetry Collector offers a Kubernetes attributes processor that enrich APM telemetry with Kubernetes metadata.

1. You need to define an environment variable in your deployment manifest.

2. Adjust the configuration of the Collector to retrieve the appropriate Kubernetes metadata using this APM environment variable.

As a result, all the APM metrics and entities will include Kubernetes metadata thanks to the K8sattributes processor. For more information, see how to link your OpenTelemetry applications to Kubernetes.

The following attributes are required in the APM service entity to display the Kubernetes summary page:

• k8s.cluster.name
• k8s.deployment.name
• k8s.namespace.name