Resource Utilization
Overview
The Resource Utilization page in Condense provides a real-time view of Kafka cluster performance and broker-level resource usage. This dashboard helps administrators monitor key operational metrics across memory, CPU, and storage, as well as request and message throughput across the cluster.
The page is accessed from: Kafka Management → Resource Utilization
Cluster Summary
At the top of the page, Condense displays an aggregated cluster status panel with the following key indicators:
Broker Status
Number of brokers online versus the total expected in the cluster.
Online Replicas
Total number of replicas currently in sync and online across topics.
Fetch Rate (kps)
The rate of fetch requests per second (in thousands) handled by the cluster.
Incoming Messages Rate (wr/s)
The number of messages written into the cluster per second.
Incoming Byte Rate (kB/s)
The volume of incoming data in kilobytes per second.
Outgoing Byte Rate (B/s)
The volume of data sent from the cluster in bytes per second.
Produce Request Rate (kps)
The rate of produce requests per second (in thousands) received by brokers.
Cloud Region
The deployment region for the Kafka cluster (if applicable).
This panel provides a consolidated health snapshot of the Kafka cluster.
Broker-Level Resource Metrics
Below the cluster summary, the dashboard displays per-broker utilization charts. Each broker in the cluster is identified with a unique label (e.g., my-cluster-kafka-3, my-cluster-kafka-4, my-cluster-kafka-5). Metrics are color-coded for easy distinction.
Memory Usage (in GB)
Shows real-time memory consumption per broker.
Values are expressed in gigabytes.
Allows identification of brokers under disproportionate memory load.
CPU Usage
Displays CPU utilization per broker as a ratio (0–1 scale).
Tracks processing load trends across brokers.
Useful for detecting uneven request distribution or overloaded nodes.
Storage Usage (in GB)
Indicates disk space usage per broker in gigabytes.
Helps monitor storage capacity and anticipate the need for scaling or partition reassignment.
Field Reference
Memory Usage
Allocated and used memory
GB
Broker-level
CPU Usage
Processor utilization ratio
Value between 0–1
Broker-level
Storage Usage
Disk space usage
GB
Broker-level
Usage Notes
All metrics are collected from Kafka brokers and displayed in real time.
Historical data can be reviewed by adjusting the time interval controls (default: minute-level sampling).
Resource imbalance across brokers may indicate:
Uneven partition distribution
Overloaded producers or consumers targeting specific brokers
Hardware performance differences
Related Sections
Last updated
Was this helpful?