The Kubernetes community has formed a Workload Conformance Working Group to write a shared standard for how workloads behave at runtime, amid a growing variety of workloads, including AI.
Broadcom's Nabarun Pal, writing on the Kubernetes blog on behalf of the working group, said "the working group is writing a shared standard for how workloads behave at runtime on Kubernetes." The effort, announced October 27, is intended to complement the Kubernetes AI Conformance test, which already determines whether a cluster can run AI or machine learning workloads. No equivalent standard exists for workloads generally, leaving teams to rely on workarounds, manual exploration, or monitoring for signs such as node drain or permission issues. The working group is seeking community input; its proposal provides more information, and interested parties can join a Slack channel or attend biweekly meetings to get involved.
The announcement comes as the Kubernetes ecosystem prepares for KubeCon + CloudNativeCon North America, November 9-12 in Salt Lake City, Utah. In other developments this week, the OpenTelemetry project's Transformation Language reached v1, stable for traces, metrics and logs and included in Collector Contrib v0.164.0 onward. Grafana Labs' Tyler Helmuth said OTTL statements keep their syntax and standard functions backward compatible. At KubeCon, AWS senior solutions architect Ratnopam Chakrabarti will present a joint Nvidia-AWS session on agent-driven GPU co-location, which he said will "demonstrate an agent pipeline that profiles workloads and handles model colocation on GPUs, all in real time." A new guidebook, Kubernetes in 2027: Building Platforms for AI, Autonomy, and Scale, published by The New Stack in partnership with HPE, covers how to stress-test infrastructure before launch and meet the demands of agentic AI.
Analysis
Why This Matters
- A shared runtime conformance standard would give teams a way to verify that their Kubernetes clusters behave predictably for workloads, rather than relying on manual exploration, workarounds, or monitoring for signs like node drain or permission issues.
- The working group's work is framed against the rapid spread of AI and agentic workloads, whose bursty, unpredictable behaviour adds new pressure on cluster infrastructure.
- The standard is being developed with community input, so teams running diverse workloads have a say in what conformance means as the ecosystem evolves.
Background
Kubernetes is the open-source container orchestration platform widely used to run cloud-native applications and, increasingly, AI and machine learning workloads. The new working group extends the project's existing Kubernetes AI Conformance test, which already checks whether a cluster can run AI or ML workloads, into a broader standard covering workloads generally.
Key Perspectives
Broadcom's Nabarun Pal, writing for the working group: The group is writing a shared standard for how workloads behave at runtime on Kubernetes, and it needs community input.
Grafana Labs' Tyler Helmuth: OTTL statements keep their syntax and standard functions backward compatible.
What to Watch
- KubeCon + CloudNativeCon North America runs November 9-12 in Salt Lake City, where the working group and related AI-on-Kubernetes sessions will be in focus.
- Whether the working group's proposal gathers enough community momentum in its biweekly meetings and Slack channel to produce a draft standard.
- Adoption of OpenTelemetry Transformation Language v1 in Collector Contrib v0.164.0 and onward, and whether the backward-compatibility promise holds as the language evolves.
Checked
12 claims in this story checked against 1 source: 11 supported.
Show each claim
- Partly supportedBroadcom’s Nabarun Pal said, “the working group is writing a shared standard for how workloads behave at runtime on Kubernetes.”The New Stack: “According to Broadcom’s Nabarun Pal, writing on behalf of WG Workload Conformance on the Kubernetes blog, “The working group is writing a shared standard for how workloads behave at runtime on Kubernetes.””
- SupportedGrafana Labs’ Tyler Helmuth said OTTL statements keep their syntax and standard functions backward compatible.The New Stack: ““If you write OTTL statements, the syntax and the standard functions stay backward compatible,” says Helmuth.”
- SupportedAWS senior solutions architect Ratnopam Chakrabarti said his KubeCon session would “demonstrate an agent pipeline that profiles workloads and handles model colocation on GPUs, all in real time.”The New Stack: “He says his KubeCon session will “demonstrate an agent pipeline that profiles workloads and handles model colocation on GPUs, all in real time.””
- SupportedThe Kubernetes community has formed a Workload Conformance Working Group to standardise workload runtime behaviour.The New Stack: “on Tuesday, October 27, the Kubernetes community announced a dedicated Workload Conformance Working Group. According to Broadcom’s Nabarun Pal, writing on behalf of WG Workload Conformance on the Kubernetes blog, “The working group is writing a shared standard for how workloads behave at runtime on Kubernetes.””
- SupportedA shared standard aims to fill a gap as workload types expand and AI adds more variety.The New Stack: “Workload types have been expanding for years. And AI undoubtedly adds more variety.”
- SupportedThe working group was announced on October 27.The New Stack: “on Tuesday, October 27, the Kubernetes community announced a dedicated Workload Conformance Working Group.”
- SupportedThe Kubernetes AI Conformance test determines whether a cluster can run AI or machine-learning workloads.The New Stack: “The Kubernetes AI Conformance test already determines whether a cluster can run AI or machine learning (ML) workloads.”
- SupportedNo equivalent standard currently exists for workloads generally, leaving teams to rely on workarounds, manual exploration or monitoring for issues such as node drain or permission problems.The New Stack: “Yet none currently exists for workloads, requiring workarounds, manual exploration, or monitoring for signs like node drain or permission issues.”
- SupportedKubeCon + CloudNativeCon North America is November 9–12 in Salt Lake City, Utah.The New Stack: “as we head into KubeCon + CloudNativeCon NA, November 9-12 in Salt Lake City, Utah.”
- SupportedOpenTelemetry Transformation Language reached v1, is stable for traces, metrics and logs, and is included in Collector Contrib v0.164.0 onward.The New Stack: “the OpenTelemetry Transformation Language (OTTL) has reached v1. It’s stable for traces, metrics, and logs, and is included within Collector Contrib v0.164.0 and onward.”
- SupportedChakrabarti will present a joint Nvidia–AWS session at KubeCon on agent-driven GPU co-location.The New Stack: “Ratnopam Chakrabarti, a senior solutions architect at AWS, took to X to share an early preview of his upcoming joint Nvidia–AWS session, which will seemingly answer this question for KubeCon attendees.”
- SupportedThe New Stack published the guidebook Kubernetes in 2027: Building Platforms for AI, Autonomy, and Scale in partnership with HPE; it covers stress-testing infrastructure before launch and meeting the demands of agentic AI.The New Stack: “In partnership with HPE, The New Stack has published a helpful book containing ten checks for teams running AI workloads on Kubernetes. The book, Kubernetes in 2027: Building Platforms for AI, Autonomy, and Scale, is available for download now. Written by contributing writer Chris J. Preimesberger, the guide spans the full lifecycle, covering how to stress-test infrastructure before launch, how to meet the nuanced demands of agentic AI”
