KubeCon + CloudNativeCon India 2025

Imagine walking up to a self-service kiosk at an airport, hospital, or hotel, speaking in your native language, and instantly hearing a real-time translation—without cloud delays or privacy risks. Traditional speech translation relies on heavy cloud compute, but what if it could run directly on low-power edge devices like Raspberry Pi or even phones?
This session demonstrates how Kubernetes (k0s) enables AI-powered multilingual speech translation at the edge, eliminating latency, cloud dependency, and high operational costs. With a single k0s control plane, running outside of edge, managing hundreds of kiosks, AI models are deployed, updated, and scaled seamlessly. We’ll showcase a live demo of real-time speech translation running on edge devices, proving how edge-native AI can revolutionize automated customer interactions.
AI-powered customer service—automated, private, and built for scale.

Running a homelab means running a diverse set of applications like NextCloud, readeck.org, usememos.com.

They are written in many languages and frameworks and in general lack a good way to monitor them. The old way to monitor them involved having a basic blackbox_exporter enabled synthetic monitoring. Even instrumented, each application has a different set of metrics, leading to complex alerting and dashboards.

Grafana Beyla, powered by eBPF and OpenTelemetry, combined with Prometheus promises to generate first class observability signals for services regardless of the language and frameworks used. We will walk through deploying Beyla for a homelab, and the challenges in having good observability even with Beyla.

We will walk through:

1. Deploying Beyla in your environment with Prometheus
2. Setting up dashboards for the data
3. Robust alerting over the signals, especially in low, sporadic traffic (homelab) scenarios.
4. The missing signals from eBPF and how you can fill those gaps

One of the significant challenges faced by Kubernetes-based applications is that performance issues often only manifest in production environments, making them difficult to reproduce in development or staging settings. Imagine if we could "time travel" and recreate real-world production conditions within a controlled environment.

This session explores how a leading global bank (200M customers in 150+ countries) successfully recreated production workloads in controlled settings. By leveraging synthetic data generation, trace playback, and workload simulation, we tackled most performance issues without compromising sensitive data. This approach not only enhanced system reliability and reduced downtime but also improved the bank's ROI by decreasing operational costs by 30% and increasing transaction efficiency by 25%, resulting in significant annual savings.

Join us as we unravel the secrets of "time travel" to replicate production performance and resolve issues effectively.

We built an enterprise-wide observability platform using CNCF Open Source tools (Grafana, Prometheus, Tempo) to address APM needs for Ops, AppDev, CloudFinOps, and Management. This self-service, automated platform offers dashboards, alerts, and notifications.

Challenges:

1️⃣ Flexibility – Observing diverse tech stacks (Java, Node, Python, Go) with standard & custom metrics and tracing.
2️⃣ Scaling – Observing workloads across multiple Kubernetes clusters and regions.
3️⃣ Cost – Fully managed solutions like Dynatrace were expensive.

Solution:

Hub & Spoke Model: Centralized tooling observing global workloads.
✅ Metrics: AWS Managed Prometheus for metrics, self-hosted Grafana for dashboards and alerts.
✅ Traces: Self-hosted Grafana Tempo with S3 storage for cost optimization.
✅ MAC: Fully automated infrastructure with AWS CDK and TypeScript for self-service configuration.

Unlock edge observability at scale. Tackle distributed node chaos with robust monitoring, logging, and security. Dive into architectural patterns for aggregating data across challenging networks. Discover how to build resilient, multi-tenant observability for your edge deployments.
Presentation leverages a real-world case study to illustrate the architectural design patterns and best practices for building scalable edge observability, emphasizing the application of core cloud-native principles and technologies

The Tenant Exporter enhances observability in multi-tenant Kubernetes clusters by delivering curated namespace metrics for GPU workloads at Adobe. Built on Prometheus, it exposes metrics like ingress requests, container CPU/memory, GPU utilization, and quotas. Users can select metrics via a ConfigMap, with quotas managing load.

The architecture features a Prometheus instance for metric collection, an Nginx proxy for load balancing, and secure authentication using prom-label-proxy with kube-rbac-proxy. A live demo will showcase configuring self-service metrics for a GPU namespace, integrated with Prometheus.

Attendees will learn to deploy this system, manage quotas, and scale metric delivery across clusters. Drawing from Adobe's experience with thousands of namespaces, this session offers best practices for observability using CNCF tools to reduce overhead and improve insights. Join us to elevate your Kubernetes observability strategy and optimize GPU workload management.