Cluster Partition Distribution Analysis #962
Description
Summary
Add a feature to Karafka Web UI that provides visibility into partition distribution across cluster brokers, helping users identify load imbalances and capacity issues.
Problem
When operating Kafka clusters, uneven partition distribution across brokers can lead to:
- Overloaded nodes running at high CPU utilization while others are underutilized
- Increased risk of cascading failures when an overloaded node goes down
- Difficulty identifying when rebalancing is needed
- Lack of visibility into which nodes are handling disproportionate load
Currently, users need to rely on external monitoring tools or manual inspection to understand their cluster's partition distribution.
Proposed Solution
Add a cluster analysis view in the Web UI that displays:
- Broker Load Overview - Visual representation of partition count per broker
- Imbalance Detection - Highlight overloaded and underloaded nodes
- Partition Assignment Table - Show which partitions are assigned to which brokers
- Replica Distribution - Display leader vs follower replica distribution per broker
Use Cases
- Identifying capacity bottlenecks before they cause failures
- Planning horizontal scaling decisions (e.g., when to add more brokers)
- Validating partition rebalancing operations
- Understanding cluster health at a glance
Additional Context
This feature would complement the existing Admin Replication API (Karafka::Admin) which already provides tools for generating rebalance assignments. The Web UI would provide the visibility layer to inform rebalancing
decisions.