Skip to main content
Zerops provides a fully managed Apache Kafka messaging platform with automated scaling and zero infrastructure overhead, letting developers focus entirely on development.

​
Supported Versions

Zerops currently supports Kafka version 3.9. When importing a service, use version format:
  • kafka@3.9

​
Service Configuration

Our Kafka implementation features optimized default settings designed for common use cases.

​
Key Configuration

  • Client Connections: Data brokers available on port 9092
  • Authentication: Secure SASL PLAIN with automatically generated credentials
  • Data Persistence: Topic data stored indefinitely (no time or size limit)
  • Performance: Optimized settings for reliability and throughput

​
Resource Allocation

Zerops automatically allocates resources to your Kafka service based on demand:
  • Memory: Up to 40GB RAM for high-performance message processing
  • Storage: Up to 250GB for persistent storage of messages and logs
  • Auto-scaling: Resources scale up and down automatically based on workload

​
Deployment Modes

Deployment mode is selected during service creation and cannot be changed later.

​
High-Availability (HA) Setup

The recommended solution for production workloads and mission-critical data:
  • Creates a multi-node Kafka cluster with 3 broker nodes
  • Configures 6 partitions across the cluster
  • Implements replication factor of 3 (each broker node has a copy of each partition)
  • Default topic replication is also 3 (overridable by user application)
  • Zerops automatically attempts to repair the cluster and data replication in case of a node failure

​
Single Node Instance

Suitable for development and testing environments:
  • Consists of 1 broker node
  • Configures 3 partitions
  • No data replication
  • Lower resource requirements
Use for development or non-critical data only, as data loss may occur due to container volatility.

​
Authentication Management

Authentication credentials are automatically generated and managed by the platform using SASL PLAIN authentication. Access your credentials through:
  • The service access details in the Zerops GUI
  • Environment variables in your service configuration:
    • user - Username for authentication
    • password - Generated secure password
    • port - Kafka port (value: 9092)

​
Client Access

Client implementations differ. Please refer to your chosen client’s configuration manual for specific details.

​
Seed Broker Connection

Connect to the Kafka cluster using the “seed” (or “bootstrap”) broker server: 
<hostname>:9092

​
Specific Broker Access

To access a single specific broker or a list of all/some brokers: 
node-stable-1.db.<hostname>.zerops:9092,node-stable-2.db.<hostname>.zerops:9092,...

​
Connection Examples

​
Node.js (KafkaJS)

import { Kafka } from 'kafkajs';

const kafka = new Kafka({
  clientId: 'my-app',
  brokers: [`${process.env.hostname}:9092`],
  sasl: {
    mechanism: 'plain',
    username: process.env.user,
    password: process.env.password
  }
});

const producer = kafka.producer();
await producer.connect();

await producer.send({
  topic: 'test-topic',
  messages: [
    { value: 'Hello Kafka!' }
  ]
});

​
Python (kafka-python)

from kafka import KafkaProducer
import os

producer = KafkaProducer(
    bootstrap_servers=[f"{os.environ['hostname']}:9092"],
    security_protocol='SASL_PLAINTEXT',
    sasl_mechanism='PLAIN',
    sasl_plain_username=os.environ['user'],
    sasl_plain_password=os.environ['password']
)

producer.send('test-topic', b'Hello Kafka!')
producer.flush()

​
Use Cases

Kafka excels in:
  • Event Streaming - Real-time data pipelines and event-driven architectures
  • Message Queuing - Decouple services with reliable message delivery
  • Log Aggregation - Centralize logs from multiple services
  • Stream Processing - Process data in real-time with Kafka Streams
  • Metrics Collection - Aggregate and analyze application metrics
  • Microservices Communication - Reliable inter-service messaging

​
Best Practices

​
Production Workloads

  • Use HA mode for all production deployments
  • Configure proper retention policies for your topics based on your data requirements
  • Monitor consumer lag to ensure messages are being processed efficiently
  • Use consumer groups to distribute processing load

​
Development Environments

  • Single node instances are suitable for development and testing
  • Be aware of potential data loss in non-HA deployments
  • Consider using smaller message sizes during development to reduce resource usage

​
Performance

  • Batch messages when possible to improve throughput
  • Use appropriate partition counts for your workload
  • Monitor broker metrics and consumer lag
  • Implement proper error handling and retry logic

​
Data Management

  • Set appropriate retention policies per topic
  • Monitor disk usage and plan for growth
  • Use compacted topics for state management
  • Implement proper key distribution for balanced partitions

​
Support

For advanced configurations or custom requirements:

Build docs developers (and LLMs) love

Get started for freeTalk to us