Kubernetes: 48% of Users Struggle With Tool Choice

Maybe the Kubernetes ecosystem is getting too complicated for its own good.

Nearly half of Kubernetes users surveyed in a new report from Spectro Cloud said they are having problems choosing and validating which infrastructure components to use within their production environments.

The main culprit: the maturity of Kubernetes.

There are simply too many choices for organizations to handle, judging by the survey participants’ responses. In the new report, 48% said they find it difficult to decide which stack components to use from the broad cloud native ecosystem. That figure has shot up from 29% who said the same in Spectro Cloud’s 2023 report.

“The Spectro Cloud 2024 State of Production Kubernetes” report is based on a survey completed in April. All of the 416 respondents work at an organization with at least 250 employees, are directly involved in their organization’s use of Kubernetes, and have at least one Kubernetes cluster in production.

Among the pain points cited by survey participants:

• Bigger deployments. Fifty-seven percent of survey participants reported having more than 20 clusters in production, up from approximately 35% who replied similarly when this same question was asked in Spectro Cloud’s 2022 report.
• That’s in comparison to 39% who said the same, in the 2023 Cloud Native Computing Foundation survey. (This statistic is only from survey participants who worked at companies with at least 500 employees.)
• Increased complexity. With the maturity of the Kubernetes market, we’ve seen an increase in the number of workloads running in Kubernetes, which are often other cloud  native elements supporting things like observability, CI/CD and service mesh. Fifty-seven percent of respondents said their Kubernetes’ infrastructure consists of more than 11 distinct software elements, up from 42% who told Spectro Cloud the same in 2022.

Complexity Causes Pain

With complexity comes headaches: Not only which tools to choose, but how to ensure they all play nicely together.

Twenty-seven percent of survey participants said the interoperability of their stack’s elements regularly causes issues with production Kubernetes clusters. That’s up from just 11% who said the same in the 2022 Spectro Cloud report.

Such interoperability issues are common: Only 26% said they rarely suffer from such issues, which means about three-quarters of production deployments have to deal with problems more than they should.

• A new set of challenges has arisen for organizations running Kubernetes in production. In addition to the difficulty survey participants report in choosing the tools they need, configuration drift (45% cited it as a challenge, up from 33% in the 2023 Spectro Cloud report) and struggles to protect against security breaches (43%, up from 26%) are the other top pain points.
• Mergers, acquisitions and companies going out of business may address some challenges. Seventy percent of respondents said they think the cloud native ecosystem is poised for market consolidation.

Adopters of Platform Engineering Suffer Fewer Issues

Platform engineering has emerged as a solution to the problems of too much complexity and too many tooling choices when running distributed systems on Kubernetes. The Spectro Cloud survey indicated the practice may be alleviating some of those issues.

• Of the 70% of organizations that have adopted platform engineering, less than half of them feel strongly that it has been fully adopted.
• Only 22% of organizations that have adopted platform engineering regularly suffer issues running production Kubernetes clusters, while 40% of non-platform engineering adopters experience those issues often.

AI and Edge Use On the Rise

Kubernetes looks likely to play a big role in the ongoing AI revolution. Sixty-eight percent of survey participants said Kubernetes infrastructure is essential in order to fully take advantage of AI in their application workloads.

Other survey results related to Kubernetes and AI:

• The most common use of AI with Kubernetes are AI assistants that help manage Kubernetes environments (44% reported this use case) and running production AI workloads on cloud or on-premises. In addition, 32% reported running production AI workloads at the edge using Kubernetes.
• Full-scale use of edge-based Kubernetes for production has risen, from 7% who reported that use in 2023 to 27% in 2024. Counting partial deployments, 38% of organizations are using Kubernetes at the edge. That means that almost three-quarters of edge Kubernetes adopters are already deploying AI workloads.