Principle:Risingwavelabs Risingwave Cluster Monitoring

Knowledge Sources	RisingWave Docs RisingWave
Domains	Observability, Operations
Last Updated	2026-02-09 07:00 GMT

Overview

An observability mechanism that provides real-time visibility into streaming job status, fragment execution, and cluster health through web dashboards and APIs.

Description

Cluster Monitoring enables operators and developers to understand the runtime behavior of a streaming database cluster. It encompasses two complementary approaches:

Built-in Dashboard: A web application served by the RisingWave meta node that displays streaming fragments, relation dependency graphs, streaming job status, and cluster information via a REST API.
Metrics Dashboards: Grafana dashboards powered by Prometheus metrics that show CPU, memory, network, storage, streaming throughput, barrier latency, and compaction statistics.

Together, these provide both logical visibility (what streaming jobs are running, what their dependency graph looks like) and physical visibility (resource utilization, throughput, latency).

Usage

Use cluster monitoring when:

Operating a RisingWave cluster in production
Debugging streaming job performance issues
Monitoring resource utilization for capacity planning
Validating that streaming pipelines are processing data correctly

Theoretical Basis

Streaming database monitoring follows a layered observability model:

Layer 1: Metrics (Prometheus)
    - Counters, gauges, histograms for all components
    - Scrape interval: 15 seconds

Layer 2: Dashboard API (Meta Node REST)
    - Fragment graph topology
    - Streaming job status
    - Relation catalog

Layer 3: Visualization (Grafana + Built-in Dashboard)
    - Time-series panels for metrics
    - Interactive fragment/dependency graphs

Related Pages

Implemented By

Implementation:Risingwavelabs_Risingwave_Dashboard_Streaming_API

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment