Google has taken another step in streamlining enterprise use of Kubernetes in a production environment by launching a fully automated pilot to steer those efforts. The next move may need to be a way to steer customers through Google’s various Kubernetes options.

The new Google Kubernetes Engine (GKE) Autopilot option is designed to manage the infrastructure needs of running Kubernetes. It brings together Google’s Site Reliability Engineering (SRE) efforts with GKE for that higher degree of automation and control.

The basics of Autopilot is that it handles all of the management of cluster infrastructure, the control plane, and nodes that are typical of day-two operations. This leaves customers to focus on their workloads.

A Google spokesperson explained that Autopilot uses GKE hardening guidelines and security best practices. This includes using GCP security features like its Shielded GKE Nodes that it launched in mid-2019, Workload Identity, and Binary Authorization.

“By locking down individual Kubernetes nodes, GKE Autopilot further helps reduce the cluster's attack surface, and minimize ongoing security configuration mistakes,” a company spokesperson noted in a statement to SDxCentral.

Drew Bradstock, GKE group product manager, added in a blog post that Autopilot also “supports maintenance windows and a pod disruption budget for maintenance flexibility, and includes for the first time a [service level agreement] on pods in addition to the SLA on hosts and the control plane.”

The platform also supports third-party options. That initial support is limited to logging and monitoring from DataDog and continuous integration, continuous delivery (CI/CD) from GitLab.

All of this management does come at a cost. Autopilot clusters charge a flat fee of $0.10 per hour per cluster for each cluster after a free tier. There is also a charge for the CPU, memory, and ephemeral storage compute resources that are provisioned for the managed pods. Google does counter that by stating the managed service does reduce the chance for a customer to be charged for unused resources.

The Google spokesperson also explained that it will continue to offer its GKE Standard edition that remains targeted at enterprises that want “complete control over all aspects of their cluster configurations.” And that GKE Standard product includes all of the features that were part of the previously launched GKE Advanced option that it rolled out in mid-2019. That service had a similar market target to Autopilot in looking to snare enterprises willing to pay more for guaranteed quality.

The Autopilot option falls into what is becoming a wide selection of managed Kubernetes offerings from where the container management platform was initially birthed.

First, there is Google’s Anthos platform. Anthos is Google’s fully managed Kubernetes-based platform that allows users to manage their data and applications in an on-premises environment or across cloud platforms from rivals like Amazon Web Services (AWS) and Microsoft. It was initially announced at Google’s Cloud Next event in 2018 — when it was labeled Cloud Services Platform — and formally launched in 2019.

The Google spokesperson explained that “GKE is a fundamental part of Anthos as we empower fleet management in GCP and extend GKE to other clouds or on-premises environments.” However, the interaction with Autopilot is limited at this point as Autopilot-managed clusters have not yet been certified to run with the Anthos control plane.

To add a bit to the confusion, Google provided a customer testimonial that noted “GKE Autopilot is the real serverless [Kubernetes] platform that we’ve been waiting for.” However, Google already offers its Cloud Run platform that is a managed serverless platform that runs on GKE and is built on the Knative open API runtime environment.

A Google spokesperson noted that while both services leverage Google Cloud capabilities like identity management, logging, monitoring, and app deployment, Cloud Run is targeted more at application developers that build event-driven applications. And then there is Google’s Cloud Functions platform that is its “serverless, event-driven platform for developers who wish to trigger code based on external events.” Clear?

Despite what appears to be a glut of options, Google has been lauded for its managed Kubernetes offerings. Forrester Research last year placed a “leaders” tag on Google’s multi-cloud container development platforms, with a specific call out on the integrated approach to container management provided by the Anthos platform.