> AI agents: this is one page from Mammoth Analytics documentation. The index of all pages as Markdown is https://mammoth.io/docs/llms.txt. Append `.md` to any docs URL, or send `Accept: text/markdown`, to get Markdown.

# Apache Kafka connector

> Connect Mammoth to Apache Kafka with bootstrap servers and a security protocol, read messages from a topic, and save them as a dataset.

This page shows how to connect Mammoth to an Apache Kafka cluster and read messages from a topic into a dataset.

Which connectors you can add depends on your plan.

## What you need before connecting Kafka

- The cluster's bootstrap servers, reachable from the internet
- Its security protocol, and a SASL username and password if it uses SASL
- Mammoth's IP address allowed through your firewall. The connection form shows the address. See [Network requirements](https://mammoth.io/docs/learn/connectors/network-requirements/).

## Connect Apache Kafka

1. Open the **Connectors** launcher and click **Apache Kafka** (see [Add a connector](https://mammoth.io/docs/learn/connectors/connector-workspace/)).
2. Enter the connection details:

| Field | What to enter |
| --- | --- |
| **Bootstrap Servers** | The broker addresses, such as `broker1:9092` |
| **Security Protocol** | Default `PLAINTEXT`. Use the protocol your cluster requires |
| **Username (SASL only)** | SASL username, if your cluster uses SASL |
| **Password (SASL only)** | SASL password, if your cluster uses SASL |
| **Display Name** | Optional label for this connection in Mammoth |



## Choose what to read from Kafka

The Kafka connector's own setup form asks:

- **Topic**: the topic to read, picked from the topics Mammoth finds on the cluster
- **Consumer Group ID**: default `mammoth-consumer`
- **Offset Reset**: **Earliest (all messages)** (default) or **Latest (new messages only)**
- **Max Messages Per Run**: default `10000`

In the connector workspace, you can tell the chat which topic you want.

## Kafka update methods

After you save, a Kafka dataset's **Update method** offers two options: **Replace all data** or **Add to existing data**. Kafka refreshes don't use a key column.

## Troubleshoot Kafka connections

- **Connection fails:** check that Mammoth's IP address can reach every bootstrap server, and that **Security Protocol** and the SASL credentials match the cluster's settings.

---
Source: https://mammoth.io/docs/learn/connectors/kafka-connector.md · Updated: 2026-11-03