# Confluent HTTP Sink

The easiest way to ingest data from Confluent Cloud based kafka pipelines is using their HTTP Sink.

# Prerequisites

To set up this integration you'll need to first:

Set up an ingest using the Array of JSON Objects schema.
Create an API Key to use with this integration with Write permissions. We'd recommend creating a dedicated Org-Wide API Key.
Follow the instructions here to install the HTTP Sink connector on Confluent Cloud

The guide and examples below assume you're using Confluent Cloud. See the docs here for local installations.

First, create a new HTTP Sink and select the source topic(s).

Choose an appropriate authentication method. These credentials need access to the topic, but will not be used by Aggregations.io.

Set the following options:

HTTP URL should be your POST URL, something like https://ingest.aggregations.io/[ID] (find it here)
Choose BASIC for the Endpoint Authentication Type
The Auth username should be blank.
The Auth password should be your API key.

Choose the relevant Input record value format for your data.

Click Show advanced configuration to reveal all the options you'll need.

Important settings:

Request Body Format should be set to json
Batch max size should be set ~ 250. Depending on the average size of your payload, you may want to increase or decrease this to ensure your payloads do not exceed 1MB
Batch json as array should be true
Retry related settings can be adjusted to your preference for robustness and resiliency.

Select the number of tasks that makes sense for your throughput and setup.

Launch the connector. Confluent Cloud will by default create an error & DLQ topic for validating everything works as expected.

See the Ingest Docs for common errors ingesting events.

From the Sink Docs:

The HTTP Sink connector does not batch requests for messages containing Kafka header values that are different.

Confirm your Kafka records have the same key.