Insights into Kubernetes-Powered Applications

Automatic, Dynamic, Full Stack Performance Management of K8S-powered Applications

Today, we announced enhanced Kubernetes monitoring support! This is an exciting moment as it brings our existing Application Performance Management to a whole new level, providing developers and operators working with Kubernetes the necessary insights not just on a resource level but more importantly on an application service level.

As this week’s KubeCon shows, Kubernetes is everywhere. It is already seen as the Cloud Operating System for Distributed Systems, as it allows teams to autonomously build and release software in a rapid fashion. These teams are often cross-functional, and handling this new cloud operating system is becoming a standard capability. Kubernetes is core to many teams and many already can’t live without it.

Essential Cluster, Deployment and Pod Metrics & Alerting

With today’s Kubernetes release we are providing Instana users the essential insights into their system. We monitor all critical components, give a comprehensive overview, and provide intelligent summaries and alerts. Here are some examples:

The Kubernetes Cluster Dashboard provides a comprehensive overview on what’s happening across the whole system. It shows: which nodes are involved, how they are doing resource-wise and how pods are scheduled. Trend graphs on the core KPIs give a quick overview and let the user identify potential problems.

Monitoring Kubernetes Cluster Dashboard

For many developers, deployments and the associated replicas are key metrics to keep an eye on. We are showing these metrics on a cluster level and in details on the Kubernetes Deployment Dashboards. The dashboard also gives deeper insights into how the pods and specific rollouts are doing.

Monitoring Kubernetes Deployment Dashboard

Alerts can be created on all important key metrics and changes. We monitor all important conditions, correlate these to other events and alert on these if they have a critical impact:

Monitoring Kubernetes Alert Screen

This is just a small subset of our new Kubernetes capabilities. Please head over to our Kubernetes Management Documentation to learn more about metrics around cluster components and how a setup is almost zero config.

But before you head over let’s talk a minute about your key asset: Your application!

Focus On Your Application

Monitoring your Kubernetes resources and building great distributed systems is certainly a fun and exciting challenge. It’s easy to get caught up with these new capabilities we now have at hand and concentrate on the technical details. But it is important to remember that in the end containers are just a means to an end; It’s all about the application or services that we provide to our internal or external customers. Sometimes with explicit, often with implicit SLAs. In the end we need to focus on the applications we are building not the technical implementation.

Instana has always excelled at identifying service impact and showing how it correlates to upstream and downstream services, independently of how the application is build. Kubernetes in that sense is just another node in our Dynamic Graph which helps our system to understand more about your environment and provide you better insights and root cause analysis.

Monitoring Kubernetes Dynamic Graph Representation

When reasoning about your application it’s important to have an overview of all involved components. More often than not these components live outside the Kubernetes cluster. That’s why our infrastructure map shows all involved entities, including all clusters and hosts independent on where they are hosted. We have simple ways to discover and filter so you only look at the components which are currently important to your application. We feel this holistic view is crucial to identify any problems that may impact your application.

Monitoring Kubernetes Application Infrastructure Map

An even more interesting view is how the services interact from a logical point of view. Our application map shows all services including the core KPIs on throughput, latency and errors. We show their relation on upstream and downstream services and how communication flows through your application.

Monitoring Kubernetes Service Map with Data Flow

With our enhanced Kubernetes capabilities you can now see not only how your services are doing and how they interact with downstream and upstream services, but you can now correlate these to Kubernetes events. If you see some degradation on your service performance, you can easily identify the involved deployments and pods. And you can now update your service with confidence as we will inform you of any problem whether they are in Kubernetes or some other components.

If you’re running Kubernetes alread, or just in the planning stages, please get in touch with Instana. If you happen to be at Kubecon stop by our booth learn more! We look forward to discussing your Kuberenetes Application monitoring and performance needs.