Skip to content
Valkey Admin

Cluster Topology

The Cluster Topology view provides an interactive table of your Valkey cluster’s nodes, showing replication relationships and connection status at a glance.

Overview

Section titled “Overview”

Understand your cluster architecture with a structured node table that groups each primary with its replica, giving you a clear picture of your replication layout.

Cluster Topology View

Cluster Statistics

Section titled “Cluster Statistics”

At the top of the page, five summary cards display key cluster metrics:

  • Total Nodes: Total number of nodes in the cluster (primaries + replicas)
  • Cluster Memory: Memory used across the cluster, out of the total memory limit
  • Total Ops/Sec: Operations per second across the cluster
  • Cluster Hit Ratio: Share of key lookups that found the key
  • Nodes Flagged: Number of primaries with High utilization

Node List

Section titled “Node List”

Layout

Section titled “Layout”

Nodes are displayed in a table. Each primary row is followed by rows for its replicas.

Node Display

Section titled “Node Display”

Each node shows its address (e.g. 192.168.18.6:7001) and a PRIMARY or REPLICA role badge. Primary rows also show:

  • Utilization: Low, Normal or High, based on the higher of memory and CPU usage. Hover the badge for details. See Utilization Levels.
  • Memory: Memory used out of the node’s maxmemory limit (e.g. 70.81M / 100 MB), or ∞ when no limit is set
  • CPU: CPU usage
  • Ops/Sec: Operations per second
  • Hit Ratio: Share of key lookups that found the key
  • Conns: Number of connected clients

Utilization Levels

Section titled “Utilization Levels”

Memory and CPU are each assigned a level, and the badge shows the higher of the two. For example, a node at 30% memory and 90% CPU shows High.

Level Memory CPU
Low Below 70% Below 60%
Normal 70% to below 90% 60% to below 85%
High 90% or more 85% or more
  • Memory is used_memory as a share of maxmemory. If maxmemory is not set, memory is not counted: the level is based on CPU alone, the badge has a dashed border, and the node’s memory limit shows as ∞. If any node has no maxmemory, the Cluster Memory card’s limit also shows ∞.
  • CPU is the main thread’s CPU time as a share of one core, measured between two consecutive samples. When you first open the page, CPU may show — until the next refresh.
  • If neither value is available, no badge is shown.

Memory has higher thresholds than CPU because a cache is expected to run close to its memory limit, while sustained high CPU on the main thread means the node is close to saturation.

Searching and Filtering

Section titled “Searching and Filtering”

Use the search bar to filter nodes by host or port. Narrow the list further with the role and utilization filters. The count of matching nodes is shown next to the filters.

Refresh Interval

Section titled “Refresh Interval”

The page refreshes node metrics, utilization levels, and the cluster statistics every 5 seconds while it is open. Refreshing stops when you leave the page.

Node Actions

Section titled “Node Actions”

Each primary row includes action icons on the right side:

  • Power: Connect to the primary node
  • Dashboard: Go to dashboard of the node
  • Terminal: Open the Send Command interface for the node

Replication Structure

Section titled “Replication Structure”

Each primary is followed by its replica:

Primary: 192.168.18.6:7001  →  Replica: 192.168.18.6:7005
Primary: 192.168.18.6:7002  →  Replica: 192.168.18.6:7006
Primary: 192.168.18.6:7003  →  Replica: 192.168.18.6:7004

Next Steps

Section titled “Next Steps”