KubeCon EU 2023 Recap

KubeCon + CloudNativeCon Europe 2023 took place in Amsterdam, drawing over 10,000 attendees from the global Kubernetes and cloud native community. Here's the THNKBIG team's summary of the key trends, announcements, and sessions that matter for engineering organizations.

Key Themes from KubeCon EU 2023

Platform Engineering Takes Center Stage

Platform engineering — the practice of building internal developer platforms that abstract Kubernetes complexity for application teams — dominated the conversation at KubeCon EU 2023. The Backstage developer portal approach, Internal Developer Platforms (IDPs), and "paved roads" architectures appeared in dozens of talks. The community consensus: simply giving developers kubectl access isn't enough. Successful organizations build opinionated platforms with guardrails.

WebAssembly (Wasm) Enters the Kubernetes Conversation

WebAssembly workloads on Kubernetes generated significant buzz. The containerd shim for Wasm (runwasi) and projects like SpinKube and Krustlet are bringing Wasm workloads into the Kubernetes scheduler as a lightweight complement to containers. Wasm instances start in milliseconds versus seconds for containers, use far less memory, and can run on architectures where container images don't exist. It's early, but senior architects should follow this space.

AI/ML Infrastructure: Kubernetes as the AI Platform

The explosion of LLM and generative AI interest is driving significant investment in Kubernetes GPU infrastructure. Multiple sessions covered GPU scheduling, MLOps platforms on Kubernetes (Kubeflow, KServe), and the challenges of multi-tenant GPU clusters. The community is actively working on better GPU scheduling primitives — Dynamic Resource Allocation (DRA) appeared in multiple forward-looking talks as the future of GPU scheduling.

Notable Announcements

• Gateway API v1.0 release candidate announced — the successor to Ingress is approaching GA with broader vendor support than ever
• OpenTelemetry marked key component GA — the collector and tracing components reaching production-ready status signals broad readiness for enterprise adoption
• Backstage joined CNCF as an Incubating project — cementing its status as the platform engineering community's reference architecture for developer portals
• Kyverno 1.10 released with enhanced policy reporting and external data source support for more sophisticated admission policies

Standout Sessions Worth Watching

All KubeCon sessions are recorded and available free on the CNCF YouTube channel. Sessions worth watching from KubeCon EU 2023: anything from the Platform Engineering track (practical IDP implementation stories), the supply chain security sessions from TAG-Security, and the Cilium gateway API deep dive for teams evaluating Gateway API alternatives to traditional Ingress controllers.

THNKBIG at KubeCon

THNKBIG attends and sponsors KubeCon events to stay current on the latest Kubernetes and cloud native developments and to connect with the broader community. If you attended KubeCon EU 2023 or watched the sessions, we'd like to hear what caught your attention. Contact us or find us at upcoming CNCF events.

KubeCon + CloudNativeCon EU 2023 in Amsterdam underscored that Kubernetes has become default enterprise infrastructure, with conversations shifting decisively from initial adoption to scaling, operational excellence, and sustainability.

Scale, Ecosystem, and Community

Record attendance and new CNCF members such as Lumigo, HCLTech, and Hitachi highlighted accelerating enterprise adoption across traditional industries, not just web-scale tech. The CNCF ecosystem has grown to over 160 projects, spanning early-stage sandbox initiatives through to graduated projects that power production systems globally. For engineering and operations leaders, this breadth means more choice but also a stronger need for clear platform standards and curation.

Day 0: Deep-Dive Specializations

Pre-conference "Day 0" events delivered some of the most practically valuable content for platform and SRE teams:

• Observability Day Europe focused on running Prometheus, Fluentd, OpenTelemetry, and OpenMetrics in production, with concrete guidance on scaling observability stacks and managing high-cardinality metrics to control cost and complexity.
• ServiceMeshCon covered production service mesh patterns, multi-cluster traffic management, and the evolving Sidecar vs. ambient mesh discussion—critical for teams designing next-generation networking and security architectures.
• Cloud Native Security Day emphasized software supply chain security, runtime threat detection, and policy-as-code for Kubernetes, reinforcing that security must be embedded into the platform layer.
• Cloud Native Rejekts showcased highly technical, experimental talks that didn’t make the main schedule but often provided the most forward-leaning ideas for practitioners.

Observability: Prometheus and OpenTelemetry

Prometheus 2.43

Prometheus 2.43 introduced several impactful capabilities for large-scale and dynamic environments:

• Out-of-order sample ingestion: This directly addresses issues in environments where timestamps are unreliable or metrics arrive late, improving robustness in real-world distributed systems.
• Native histograms: More efficient and accurate representation of distributions, especially valuable for latency and performance metrics.
• Memory usage improvements: Better handling of high-cardinality workloads, reducing resource pressure on Prometheus servers.

For platform teams, these changes reduce operational friction and make Prometheus more suitable for complex, heterogeneous environments.

OpenTelemetry Advancements

OpenTelemetry (OTel) continued to mature as the vendor-neutral telemetry standard:

• Histogram support: Aligns with modern metrics needs and complements Prometheus’ native histogram work.
• Log bridge: Enables legacy logging systems to feed into OTel pipelines, easing gradual migration and integration.
• OTLP protocol progress: Finalization of the OTLP specification simplifies building standardized, portable telemetry pipelines.
• Convergence with Elastic ECS: Alignment with Elastic Common Schema reduces integration friction for teams using Elastic stacks alongside OTel.

Together, these updates strengthen the foundation for unified, vendor-neutral observability across metrics, logs, and traces.

Sustainability in Cloud Operations

Sustainability emerged as a substantive technical theme rather than a purely strategic talking point:

• Red Hat’s session on Kepler (Kubernetes-based Efficient Power Level Exporter) demonstrated how to use eBPF to measure per-Pod energy consumption.
• This level of granularity enables teams to understand the energy and carbon impact of specific workloads, a capability that was previously missing.
• For organizations with sustainability targets or operating under carbon pricing regimes, Kepler’s metrics support data-driven optimization of workloads, capacity planning, and placement strategies.

Emerging Directions

Conference content pointed to several key future directions for cloud-native engineering:

• Platform Engineering & Internal Developer Platforms (IDPs): Increasing focus on curated, self-service platforms that reduce cognitive load for application teams while enforcing security, compliance, and reliability standards.
• AI/ML on Kubernetes: Growing momentum around running ML training and inference workloads on Kubernetes, leveraging cloud-native patterns for reproducibility, scaling, and automation.
• WebAssembly (Wasm) as a cloud-native runtime: Wasm is emerging as a complement to containers, offering fast startup, strong isolation, and portability for certain classes of workloads.

Across these areas, core cloud-native principles—automation, declarative configuration, GitOps, and strong platform abstractions—are being applied to an increasingly diverse set of workload types.

What This Means for Engineering and Operations Leaders

• Kubernetes is now baseline infrastructure, not an experiment. The focus must shift to operational excellence, governance, and developer experience.
• Observability stacks built on Prometheus and OpenTelemetry are stabilizing as the de facto standard, with new capabilities that better support scale, cost control, and interoperability.
• Sustainability metrics are becoming first-class citizens in platform design, influencing architecture, workload placement, and optimization strategies.
• The most advanced patterns and ideas often surface in Day 0 events and Rejekts, making them essential for teams that want to stay ahead of the curve.

THNKBIG attended KubeCon EU 2023 to stay aligned with the leading edge of the CNCF ecosystem and the projects we implement for clients. Talk to our team about applying these cloud-native patterns—platform engineering, observability modernization, sustainable operations, and emerging runtimes like Wasm—to your organization’s infrastructure and delivery platforms.