Survey Surfaces Where Kubernetes Clusters Are Running Most
A survey of 125 IT professionals finds that three-quarters (75%) of respondents are running Kubernetes clusters in public clouds compared to 43% running Kubernetes in a private data center.
Conducted by Loft Labs, a provider of a platform for creating and managing virtual Kubernetes clusters, the survey finds that 43% are also employing managed Kubernetes services.
Only 19% of respondents said they are running Kubernetes across a multi-cloud computing environment. The most widely used platforms are Amazon Web Services (AWS) at 62%, followed by bare metal servers at 37%, Microsoft Azure (31%), Google Cloud Platform (29%) and VMware (15%).
The Loft Labs survey finds that 55% of respondents work for organizations that have deployed 10 to 50 Kubernetes clusters, while 25% have fewer than ten. The most widely used tools are kubctl (100%), followed by jq, K9s, kubectx, vCluster, Krew and DevSpace. vCluster and DevSpace were both developed by Loft Labs. Well over half of respondents (57%) ranked their Kubernetes proficiency as high.
Rahul Patwardhan, senior director for demand generation at Loft Labs, said that while more than 40 million virtual Kubernetes clusters have been created there are still a large number of workloads running on physical servers that could be more cost-effectively run on a virtual cluster. In general, virtual clusters are used most widely in pre-production environments. However, as the number of cloud-native applications being deployed continues to grow there is an opportunity to reduce costs by using virtual clusters in production environments, said Patwardhan.
Originally launched in 2021, open-source vCluster software enables IT teams to create lightweight Kubernetes clusters that run inside the namespaces of a multi-tenant cluster using either a command line interface (CLI) favored by DevOps teams or a graphical user interface (GUI) tool favored by IT administrators that are both provided by Loft Labs.
It’s not clear how much pressure IT teams are under to reduce the total cost of IT. However, there is a tendency for application development teams to want to provision physical infrastructure either in the cloud or on-premises. Upon further review, however, many of those workloads could run on shared infrastructure to both reduce costs and increase overall flexibility, noted Patwardhan.
The one certain thing is more cloud-native applications than ever will soon be deployed on Kubernetes clusters. In many cases, Kubernetes has now become the default platform for deploying modern container applications. The challenge, of course, is finding and retaining the expertise needed to manage those environments.
Theoretically, Kubernetes environments should become easier to manage as further advances in artificial intelligence (AI) are made. Most Kubernetes clusters today are managed by DevOps teams. However, it is becoming increasingly easier for IT administrators who generally lack programming expertise to manage these platforms using commercial frameworks.
Each organization will ultimately need to decide where to deploy Kubernetes clusters, however, as their number continues to increase the workflows and processes used today to manage Kubernetes clusters may not scale much further. The added dimension now being added to that debate is how many virtual clusters might soon be part of that mix.
