Что произойдёт при отказе брокера Kafka?

Лидерство партиций автоматически перейдёт к репликам на живых брокерах, клиенты переключатся штатным механизмом протокола Kafka, а упавший брокер платформа пересоздаст.

Как управлять топиками в managed Kafka?

Декларативно: набор топиков с партициями, репликами и политиками хранения — часть конфигурации сервиса, что исключает дрейф настроек и подходит для GitOps.

Совместим ли managed Kafka с Kafka Connect и стандартными клиентами?

Да, это стандартный Apache Kafka — любые клиенты и инструменты экосистемы работают без изменений.

Можно ли подключаться к Kafka снаружи облака?

Да, внешний доступ включается отдельной опцией — подключение идёт через LoadBalancer и защищено TLS.

Managed Apache Kafka in the cloud — a cluster in minutes.

Managed Apache Kafka

An Apache Kafka cluster in a few steps: three brokers with replication, declarative topic management, and traffic encryption. The service lives in your DevSpace, alongside the project’s databases and applications — build an event‑driven architecture, and leave updates, recovery, and cluster coordination to the platform.

A streaming platform for events and integrations — without the hassle of administration.

Free Trial

What you get

-1-

A cluster ready for production

several brokers with partition replication between them
-2-

Declarative topics

Partitions, replication factor, and storage policies are defined by configuration, not by scripts on the broker
-3-

Traffic encryption

TLS between cluster components and on external connections
-4-

Kafka standard protocol

All clients in the ecosystem are working: Java, Go, Python, Kafka Connect, ksqlDB
-5-

Scaling

The growth in the number of brokers and resources as the flow increases

Free Trial

How the service works

The cluster is managed by Strimzi — an open‑source Kubernetes operator for Apache Kafka and a CNCF project, the de facto standard for running Kafka on Kubernetes.

Partition replication

Each topic partition is stored in multiple copies across different brokers; if a broker fails, the leadership automatically switches to a replica, and the data flow is not interrupted.
Topics as a configuration

The topic is described declaratively: name, number of partitions, replication factor, policies (retention, compaction, and other Kafka settings). The operator brings the cluster to the description — this is convenient for GitOps.
Managed lifecycle

Strimzi performs rolling updates — brokers are restarted one by one without stopping the cluster, and it monitors configuration consistency.
TLS within the cluster — always

Inter‑broker traffic and service connections are encrypted; the cluster’s PKI is managed automatically by the operator.
Persistent volumes

Broker data is stored on the platform’s block storage; replicated storage classes are available.

Opportunities

Use cases

Event‑driven architecture

Asynchronous communication between microservices instead of fragile synchronous chains. The services themselves can run in managed Kubernetes in the same space and communicate with brokers using internal addresses — there’s no need to expose Kafka to the outside
Stream processing

metrics, clicks, telemetry, IoT: real‑time event ingestion and processing
Integration bus

Reliable data delivery between systems, including CDC replication from databases
Buffer and unloading

Smoothing out load peaks between producers and consumers

Free Trial

Как развернуть Apache Kafka в облаке — три шага

1

Add the “Kafka” service to your DevSpace: broker resources and storage size.
2

Describe the topics — partitions, replicas, and retention policies — directly in the service configuration.
3

Connect the producers and consumers using standard Kafka clients at the address from the panel.

FAQ

Try it out