Google Cloud puts its Kubernetes Engine on autopilot

Google Cloud unveiled today a new operating mode for its Kubernetes Engine (GKE), which will enable the Google's own engineers and automation software to handle most of the everyday running of a container cluster.

Google handles all Day 2 activities for the management of these clusters and their nodes, with Autopilot, as the new mode is called, while enforcing good practices for their service and security.

This new mode improves the current GKE experience, which has already handled much of the cluster infrastructure.

As Google Cloud considers this 'ordinary' experience is still possible and enables users to tailor their setups to the substance of their hearts as well as to manually supply and maintain their node resources.

Drew Bradstock, GKE Group Product Manager, said the concept behind Autopilot was to get all the tools Google previously had for GKE together together with their SRE teams who knew how to produce these clusters – something Google has done in the business for a long time.

Bradstock noted that "Autopilot stitches together auto-scaling, auto-upgrades, maintenance, Day 2 operations and — just as importantly — does it in a hardened fashion,"

"[...] What this has allowed our initial customers to do is very quickly offer a better environment for developers or dev and test, as well as production, because they can go from Day Zero and the end of that five-minute cluster creation time, and actually have Day 2 done as well."

Google Picture Credits:

Nothing improves here from a developer viewpoint, but this new mode frees teams to concentrate on the real workload and not on handling clusters in Cubernet.

With Autopilot, enterprises also benefit from Kubernetes but without any scheduled management and maintenance work.

And this is undoubtedly a development we saw when the ecosystem of Kubernetes grew.

Over all, few businesses consider Kubernetes as their true strategic differentiator.

All this inevitably comes at a flat rate of $0.10 per hour and cluster (a free GKE tier also offers billing credits of $74.40), plus of course the normal resources costs for the clusters.

Google provides a 99.95% SLA for the Autopilot cluster control plane and a 99.9% SLA for the Autopilot pods in some regions.

Autopilot for GKE follows a series of Google Cloud container-centred products, including Anthos for use in multi-cloud environments and Cloud Run, Google's serverless offer.

"[Autopilot] is really [about] bringing the automation aspects in GKE we have for running on Google Cloud, and bringing it all together in an easy-to-use package, so that if you're newer to Kubernetes, or you've got a very large fleet, it drastically reduces the amount of time, operations and even compute you need to use," said Bradstock.

And though GKE is an essential part of Anthos, this service offers Google's configuration management, service mesh and other resources more to the data center of one organization.

GKE autopilot is only available on Google Cloud, at least for now.

"On the serverless side, Cloud Run is really, really great for an opinionated development experience," said Bradstock.

"You can get on with it really quickly when you want an application that can go from scratch to scratch to scratch — not worry and manage it entirely by Google. That's highly valuable and perfect for much development. It's more about simplifying the entire platform that people are working on, when they want the Kubernetes eco-system to be leveraged, much more controlled and more controlled.