Google Tool Monitors Cloud-Hosted, On-Premises Kubernetes Clusters
The company’s Stackdriver Kubernetes Cluster Monitoring tool will make it easier to monitor cluster-based app environment for performance issues.
Organizations using the Kubernetes platform to build and deploy container applications now have a new option for monitoring the environment for performance-impacting bugs, bottlenecks and abnormal behavior.
Google this week announced the beta release of Stackdriver Kubernetes Monitoring for aggregating logs event data and other metrics from Kubernetes environments. Developers and operations teams can use the data to understand—in near real time—how an application is performing in production and to quickly identify issues that might impact performance.
“Up until now, observing a complex Kubernetes environment has required manually stitching together multiple tools and data coming from many sources, resulting in siloed views of system behavior,” Google Product Manager JD Velásquez wrote in a blog May 2.
Stackdriver Kubernetes Monitoring eliminates the need for that by enabling the integration of events, metrics, logs and metadata from across the environment. The centralized data can help speed tasks like issue identification and root cause analysis and reduce the mean time needed to resolve an issue.
In addition to enabling faster time to resolution, Stackdriver Kubernetes Monitoring enables access to logs and metrics from containers deployed across multiple environments and provides customized views for developers, operations staff and security engineers.
Developers can use the new monitoring feature to observe more closely all the clusters, services, pods, workloads and other Kubernetes objects within their application so they can understand what normal behavior looks like. Such visibility can help reduce or even eliminate the need for constant instrumentation and management of the Kubernetes infrastructure.
Site Reliability Engineers (SREs), meanwhile, can use Stackdriver Kubernetes Monitoring to centrally manage multiple Kubernetes container clusters in both public and private cloud environments. The tool gives SREs a way to quickly view the overall health of each container cluster and to drill down into the state of individual Kubernetes objects so they troubleshoot issues more effectively, Velásquez said.
Security administrators similarly can use the data available through the new tool to look for and mitigate potential security-related issues in Kubernetes container-based application environments.
Google’s new Stackdriver Kubernetes Monitoring technology works with the open-source Prometheus tool. Organizations can ingest metrics from third-party applications and custom applications using Prometheus instrumentation, and the data will work without modification in the Kubernetes monitoring environment.
Organizations will be able to use Stackdriver Kubernetes Monitoring right away, since it is preintegrated with Google’s Kubernetes Engine, Velásquez said. The tool also works with Kubernetes deployments on other public cloud platforms and on-premises infrastructure. Regardless of where containers are deployed, Stackdriver Kubernetes Monitoring will give administrators access to a centralized collection of logs and other application metrics, he noted.
Google pricing for Stackdriver Kubernetes Monitoring is based on the notion of so-called chargeable resources, or resources that are in a running or runnable state. Enterprises that are signed up for the Premier Tier are charged $8 per month for the Google cloud resources—like each App Engine, Compute Engine or Container Engine node instance—that they actually use.
Starting June 30, Google will introduce what it describes as a new and simpler pricing model designed to give organizations more control over their Stackdriver usage. The current tiered pricing system will be eliminated, and there will no longer be any premium pricing options.