Product
Features
Everything a platform team needs to run Apache Kafka® with KRaft, in one dashboard and one REST API.
Clusters
One page shows every cluster: health, brokers (including fenced and cordoned ones), the KRaft quorum, topics, partitions, under-replicated and offline partitions, and versions. Each cluster has its own page with brokers, controllers, topics, the partition distribution, consumer and share groups, configuration and features. When no broker answers, Brokerium reads the quorum straight from the controllers.
Brokerium connects with PLAINTEXT, SASL/PLAIN, SASL/SCRAM, TLS and mutual TLS, Amazon MSK IAM (static keys or the AWS default credential chain) and presets for Confluent Cloud and Aiven. Stored credentials are sealed with the server key and never returned by the API.
Controllers
- See voters, observers, the leader and how far each controller lags.
- Add and remove controllers one at a time. A new controller joins and catches up before an old one is removed.
- Migrate a static quorum (
controller.quorum.voters) to a dynamic one in place.
Rebalancing
- A deterministic planner spreads replicas, leaders and data across brokers and racks; the same input always gives the same plan.
- Balance evenly, balance leaders only, elect preferred leaders, or move replicas and leaders by hand.
- Every plan shows the distribution before and after, the number of moves and the data to move.
- Reassignments run throttled, in batches, behind the health gate, and the throttles are always removed.
Brokers
Add, drain, decommission and replace brokers, with the same node id on an empty disk or with a new id. Rolling restarts go one node at a time. Existing nodes can be adopted without touching their configuration or storage.
Upgrades
Rolling upgrades check the Java runtime and known Kafka defects (for example KAFKA-20979) before they start, upgrade one node at a time behind the health gate and finalize the metadata version as a separate, explicit step.
Data plane
Topics and their configuration, partitions, ACLs, SCRAM users, client quotas, consumer and share groups with offset resets, message browsing and producing.
Platforms
| Where Kafka runs | What Brokerium uses | What you get |
|---|---|---|
| Virtual machines | The Kafka protocol, plus brokerium-agent on each host | Everything, including node lifecycle and upgrades |
| Kubernetes with Strimzi | Strimzi custom resources through the Kubernetes API | Overview, data plane, rebalancing, node pool scaling, rolling restarts and upgrades; Strimzi owns the quorum |
| Amazon MSK, Confluent Cloud, Aiven | The Kafka protocol with the provider’s authentication | Overview and data plane, plus what the provider allows |
| Any cluster you can reach | The Kafka protocol with a service account | Overview, data plane and rebalancing |
Every scenario on this page runs as an automated end-to-end test against a lab of real Kafka clusters before a release.