Today, everyone in the industry is “all-in” on Kubernetes – and for good reason. Running workloads on top of Kubernetes is significantly valuable, not just for application teams, but for infrastructure teams as well. Beyond its high availability, Kubernetes supports hybrid-infrastructure, allowing these teams to run their workloads on different Kubernetes environments across a variety of clouds and on-premises infrastructure. Importantly, Kubernetes also enables developers to build, deliver, and update microservices-based applications flexibly, reliably, and quickly.
However, while Kubernetes can help teams monitor the health of their environments and restart failed applications, the platform has limited visibility into the internal state of those applications. When it comes to observing Kubernetes environments, your approach must be rooted in metrics, logs, and traces—and also the context in which things happen and their impact on users.
While there are several solutions you can use to monitor the state of a cluster and receive alerts when anomalies occur, Dynatrace is the only solution that offers intelligent observability into the entire application landscape. At the core of this approach is the Dynatrace AI engine, Davis®, which automatically delivers an in-depth analysis and precise root cause whenever anomalies arise.
We sat down with Dynatrace’s Alois Mayr, Senior Technical Product Manager, at Perform 2021 to learn how teams can harness the power of automation, AI, and enterprise management capabilities to tackle the complexity of large Kubernetes environments. Alois walked us through how the Dynatrace platform helps teams uncover the health and resource utilization of every workload, pod, and node within clusters, and how this collectively impacts microservices and end-user experience.
A single source of truth
One of the most important goals of an operations team is to understand whether a cluster is healthy. Dynatrace provides the insights to help teams determine this, while also uncovering a range of additional insights, including event tracking and over-commitment rate. The Dynatrace dashboard provides all the information teams need about workloads and pods, and which phases they are running, all within a consolidated dashboard view.
Dynatrace’s OneAgent automatically collects information from a variety of sources, from the nodes and pods, and within the pods, to the environment’s API. Using this information, Dynatrace detects dependencies between errors and individual services to help teams resolve issues quickly. To understand the root cause, the Dynatrace AI engine, Davis®, uses AI-driven PurePath technology to analyze the journey of an individual user request in the browser and trace all the way to the back end to see how it’s contributing to the problem, down to the line of code that was called. Further, Session Replay allows teams to visually replay each user’s experience with live session details.
Next-level application performance insights
Beyond identifying services that have been impacted, Dynatrace offers fully customizable dashboards that help teams evaluate the impact of their technical investments on meaningful business KPIs such as revenue and conversion rates. This information arms teams with an understanding of how new features are resonating with customers in a visual format.
- New logs support for Kubernetes – new integration with Fluentd enables Dynatrace to automatically capture log and event streams from Kubernetes and multicloud platforms, including AWS, GCP, Microsoft Azure, and Red Hat OpenShift. This will provide teams insights from extended log streams for enriched root-cause analysis.
- Control plane – updates to better understand control plane health, new dashboards (etcd, api-server, controller manager, kubelet).
- More dedicated dashboards – including those in development for Istio, LinkerD, Knative.
You can watch the full session and learn more about how Dynatrace is accelerating innovation with Kubernetes below.
More about Kubernetes
Learn more about the Kubernetes architecture, options for running Kubernetes across a host of environments and the adoption patterns of cloud native infrastructure and tools.