⚛️ Data Mesh Reference Architecture · Quarkus
Tutorial Setup Demos GitHub ↗
Home / Operating the mesh
Part 2 3 chapters

Operating the mesh

Running a data mesh day to day: progressive delivery and mTLS over Istio, scaling to demand and to zero with KEDA, and the observability to see all of it working.

07

Progressive delivery and mTLS

Evolving a data product's contract in the open with an Istio v1→v2 canary on a local Kubernetes cluster, and the decision to mesh selectively rather than enable sidecar injection namespace-wide.

30 minutesRead →
08

Elastic and resilient

Scaling two data products to demand, and to zero, with KEDA; the manifests and demos that exercise it; and the recoverability Kubernetes provides automatically.

30 minutesRead →
09

Observability

Metrics, distributed traces, and the live mesh view — the LGTM stack and OpenTelemetry wiring this repo installs, and a verified cross-service trace.

30 minutesRead →
← Previous part Building data products Next part → Lessons & close

Data Mesh Reference Architecture · Quarkus

A working data-mesh reference implementation in Quarkus — services across REST, gRPC, GraphQL, and Kafka, backed by Camel-on-Quarkus routes and langchain4j/MCP tooling, with contracts, a catalog, progressive delivery, autoscaling, and full observability. Includes a runnable Spring Boot twin service for a side-by-side comparison.

Reference

  • Setup & prerequisites
  • Demos & examples

Built with

  • Quarkus ↗
  • Camel Quarkus ↗
  • langchain4j ↗

Project

  • View on GitHub ↗
  • Issues ↗
  • Python sibling repo ↗
Open source · Apache 2.0 Last built 2026-10-09