Information technology monitoring startup Grafana Labs has announced a long list of updates to its platform in order to help companies to address some of the key challenges around Kubernetes monitoring, ahead of the KubeCon + CloudNativeCon Europe event that kicks off on March 19 in Paris.

New York-based Grafana Labs is the developer of Grafana Cloud, one of the most popular open-source platforms for monitoring IT infrastructure. The platform transforms telemetry and operational data from a company’s infrastructure, such as servers and databases, into graphs that can help administrators more easily uncover useful patterns. In this way, it can help IT teams detect technical issues before they cause significant problems. When it spots a potential problem, it will generate an alert, giving teams advance warning so they can take action to prevent downtime.

Grafana provides monitoring tools for various IT infrastructure platforms, and one of its most popular offerings is Kubernetes Monitoring in Grafana Cloud. It makes it simpler for companies to monitor Kubernetes clusters by allowing them to visualize and analyze key metrics related to their environments, track resource usage, and gain insights into the behavior of their applications within Kubernetes.

Kubernetes is an open-source software platform that’s used to manage large clusters of containers, which host the components of modern software applications that can run anywhere. It’s one of the most widely used and critical pieces of software for thousands of enterprises across the globe, serving as the foundation of their most critical business applications.

Feature updates

With today’s update, Grafana is making Kubernetes monitoring easier than ever to deploy and scale. Customers no longer have to use Grafana Agent or Grafana Agent Operator to manually configure their Kubernetes infrastructure data. Instead, they can use the new Grafana Kubernetes Monitoring Helm chart, which automates the transfer of metrics, logs, events, traces and cost metrics to Grafana Cloud.

A second update makes it possible for users to respond to and troubleshoot Kubernetes alerts without leaving the context of the Grafana Kubernetes Monitoring platform. Users can begin their troubleshooting operations via the “Pods in trouble” section on the home page, which takes them to an update Alerts page, where they can view graphs detailing alerts by cluster and namespace, ranked in order of alert severity. This will help users to resolve the most pressing issues faster, with lower-priority issues set aside until teams have adequate time to deal with them.

Visibility is improved too, thanks to a new feature that provides a more comprehensive overview of each cluster’s health and performance. The Grafana Kubernetes Monitoring home page now offers snapshots of critical issues, with graphs for each affected cluster, node, pod and container, regardless of which cloud provider or Kubernetes distribution is being used. There’s also a new “time picker” feature that can quickly display historical data analysis, so users can examine resource usage over specific time frames to better address inefficiencies and cost management.

In order to monitor, predict and optimize resource usage, Grafana has introduced a new “Summary Views” function for each cluster, node, workload and namespace that correlates their compute, memory and storage data to aid in performance troubleshooting and identifying underused resources. This feature also enables users to forecast future compute and memory use so they can optimize resource allocation ahead of time, leading to improved efficiency and cost-effectiveness, the company said.

To further aid with this, Grafana has introduced new cost monitoring tools that correlate resource management with cost attribution, so users can see a breakdown of costs on a per-container and per-pod basis.

Integrations and support

Today’s updates extend beyond new functionality, with Kubernetes Monitoring in Grafana Cloud now being listed in the Amazon Web Services Marketplace as an Amazon EKS add-on. AWS users can also integrate the platform with ClickHouse, InfluxDB and Presto, the company said.

In addition, the company said its open-source eBPF-based auto-instrumentation tool Beyla has gained full support for Kubernetes, allowing users to incorporate Kubernetes metadata into the telemetry it generates. Grafana Beyla makes it possible to instrument applications through a single command, without modifying the underlying code. With this update, Grafana Beyla can now better “understand” Kubernetes semantics to provide a more fine-grained selection of services to instrument.

Finally, Grafana provided an update on its ongoing support for a number of key open-source projects sponsored by the Cloud Native Computing Foundation. It explained that it is one leading contributors to both the OpenTelemetry and Prometheus projects at a time when they’re seeing accelerated adoption by enterprises. OpenTelemetry is a set of standards and tools to capture and export telemetry data from cloud-native applications and infrastructure, while Prometheus is a monitoring solution for collecting and aggregating metrics as time series data.

In its 2024 Observability Survey, Grafana found that 89% of respondents are investing in Prometheus, and 85% are investing in OpenTelemetry, with almost 40% reporting that they use both to aid in their observability operations. As such, Grafana has devoted significant resources to ensuring these projects are interoperable with its own platform.

Juraci Paixão Kröhling, principal software engineer at Grafana, said the company’s support of OpenTelemery dates back to that project’s predecessors, OpenCensus and OpenTracing. “Our investment in the project has only increased over time as we identify areas that make sense for our users,” he said. “With metrics, and more recently logging, marked as stable in OpenTelemetry, we’re seeing the project gain momentum among Grafana users as more people are coupling it with Prometheus as the backend. Because we want to be where our users are, our goal is to continue to make it easier for them to get value from OpenTelemetry.”

Image: Grafana Labs