> ## Documentation Index
> Fetch the complete documentation index at: https://docs.vortexiq.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Replicas per Node, CockroachDB

> Replicas per Node for CockroachDB clusters. Tracked live in Vortex IQ Nerve Centre. How to read it, why it matters, and how to act on it.

**Card class:** [Non-Hero](/nerve-centre/overview#card-classes-explained)  •  **Category:** [Ranges & Leases](/nerve-centre/connectors#connectors-by-type)

## At a glance

> **Replicas per Node** shows how many range replicas each node in the cluster is currently storing, compared across nodes as a bar chart. CockroachDB splits data into ranges and replicates each across multiple nodes; the allocator continuously rebalances so every node carries a roughly even share. A flat, even bar chart is a healthy, balanced cluster. A lopsided chart, one node well above or well below its peers, signals a rebalance in progress, a recovering node, or an allocation constraint that is preventing even distribution.

## What it tracks

This card reports **Replicas per Node, compared across items**: the live count of range replicas held on each node, refreshed in real time (`RT`). The data comes from CockroachDB's per-store replica counts, exposed via the `replicas` metric and the cluster status layer. It is a distribution view, so the bar chart matters more than any single number; you are reading the shape across nodes.

A balanced cluster keeps replica counts within a narrow band of each other because the allocator rebalances ranges toward even distribution. Imbalance is expected and benign in three situations: a freshly added node climbing from zero as ranges move to it, a restarted node catching up, and an active decommission draining a node toward zero. Imbalance is worth investigating when it persists with no rebalance underway, which usually means an allocation constraint (zone configuration, locality, or disk pressure on one node) is blocking even placement. There is no alert threshold on this card (`alert: -`); it is a diagnostic companion to the lease and replication cards.

Note that replica count is not the same as leaseholder count: a node can hold many replicas but few leaseholders. Leaseholders, not replicas, carry the read/write serving load, so for a hot-node diagnosis pair this with [Range Lease Balance Skew %](/nerve-centre/kpi-cards/cockroachdb/range-lease-balance-skew). For replication-safety questions pair with [Under-Replicated Ranges](/nerve-centre/kpi-cards/cockroachdb/under-replicated-ranges) and [Raft Quiescent Lag (seconds)](/nerve-centre/kpi-cards/cockroachdb/raft-quiescent-lag-seconds).

**Reconcile** against **DB Console → Metrics → Replication dashboard → "Replicas per Store"**, or run **`cockroach node status --ranges`** from the CLI for per-node replica and leaseholder counts. The raw per-store figures are also available from `crdb_internal.kv_store_status`. In CockroachDB Cloud the same chart sits under **Monitoring** in the Cloud Console.

***

### Tracked live in Vortex IQ Nerve Centre

*Replicas per Node* is one of hundreds of KPI pulses Vortex IQ tracks across CockroachDB and 70+ other ecommerce connectors. Nerve Centre runs the detection layer; Vortex Mind investigates the cause when something moves; Ask Viq lets you interrogate any number in plain English.

[Start for free](https://app.vortexiq.ai/login) or [book a demo](https://www.vortexiq.ai/contact-us) to see this metric running on your own data.