Dubbo Admin 0.7.0: From Registry Console to Cloud‑Native Control Plane

Version Overview

Apache Dubbo Admin 0.7.0 redesigns the console as a Kubernetes‑native control plane. The release shifts the product from a traditional registration‑center console to an event‑driven, cloud‑native architecture that runs inside a Kubernetes cluster.

Architecture

Compatibility with traditional registries and Kubernetes – supports classic registries (Zookeeper, Nacos) and Kubernetes Service models, providing global visualization of applications, instances, and services and enabling multi‑registry and multi‑cluster deployments.

Governance capability upgrade – a unified traffic‑governance UI; the backend automatically converts and distributes routing rules to both traditional registry resources and Kubernetes resources.

Deep observability integration – beyond metadata, the console displays real‑time call‑chain metrics, supporting Metrics and Tracing visualizations.

Kubernetes‑native design – backend rewritten in Go, tightly integrated with client-go, informers, and caches.

Domain‑Driven Design & CRDs

Dubbo Admin models the Dubbo domain entities—Application, Service, Instance, and Rule—as Kubernetes Custom Resource Definitions (CRDs). The CRD set unifies resources from traditional registries and Kubernetes, allowing a single event‑driven informer framework to list‑watch updates and cache them locally for low‑latency queries.

Core Highlights

UI Refresh

The UI is rebuilt around the Dubbo domain model, linking applications, instances, and services. Classic routing rules (dynamic, conditional, tag‑based) are presented as forms, lowering the barrier for traffic management.

Kubernetes‑Native Control Loop

Admin runs as an in‑cluster component that reconciles Dubbo instances with pod lifecycles using a Controller / Reconcile / Desired‑State pattern.

Discovery & Multi‑Registry

Native discovery implementations for Nacos and Zookeeper are integrated into the control plane. Each registry is treated as an isolated domain, enabling seamless switching, hybrid deployments, and unified management of both Kubernetes Service‑based services and traditional registry services.

Observability × Governance Closed Loop

Dubbo SDK 3.2+ embeds metric and trace instrumentation. Metrics (QPS, latency, error rate) and tracing (OTLP, Jaeger, Zipkin) are collected and visualized via Grafana panels for service‑level and method‑level trends, latency, error rates, and full call chains.

Event‑Driven Informer Layer

Leveraging client-go informer mechanisms, Admin uses a unified list‑watch model for both registry and Kubernetes resources. All resources are cached in a memory store that provides low‑latency, strongly consistent reads and high‑throughput indexing, enabling “change‑as‑response” behavior.

Observability Integration

Metrics are exposed through Prometheus; tracing data can be sent to Jaeger or Zipkin. Grafana serves as the unified visualization foundation, with built‑in panels for Dubbo call monitoring and trace analysis.

Repository

Source code: https://github.com/apache/dubbo-admin