Getting started with the Webhook sink

The webhook integration provides a way to call a predefined HTTP endpoint for each new batch of data produced by a DNA stream.

This integration is well suited for the following tasks:

• build an indexer using serverless technologies. You can invoke serverless functions such as, but not limited to, AWS Lambda, Supabase Functions, Cloudflare Functions, and Vercel Functions.
• start a background job on Inngest or other job queues.
• push messages into a message queue like Kafka or AWS SNS.
• send notifications using the Discord or Telegram API.

apibara plugins install sink-webhook

In standard mode (default), the webhook is invoked with the following payload. The cursor property contains the block number and hash (cursor) of the first block in the batch, while end_cursor contains the cursor of the last block in the batch.

type Cursor = {
  /** Block number. */
  orderKey: string;
  /** Block hash. */
  uniqueKey: string;
};

type Payload = {
  data: {
    cursor: Cursor;
    end_cursor: Cursor;
    finality: string;
    batch: object[];
  };
};

On chain reorganizations, the webhook is invoked with the following payload.

type Payload = {
  invalidate: {
    cursor: Cursor;
  };
};

In raw mode, the webhook is invoked with the data returned by the transform function. This mode is useful if you need to send data to an external service that expects the payload in a specific way. Users use raw mode to implement Telegram and Discord bots.

Notice that chain reorganizations are ignored when using raw mode.