Documentation

Architecture of the Indexer SDK

In the Aptos indexing stack, a processor indexes a specific subset of data from the blockchain and ingests the data into an external database.

Each processor follows this general flow:

1. Receive a stream of Aptos transactions
2. Extract the relevant data from the transactions
3. Transform and merge the extracted data into a coherent, standardized schema
4. Store the transformed data into a database

The Aptos Indexer SDK allows you to write and model each of your processor as a graph of independent Step’s’. Each Step has an input and an output, and the output of one Step is connected to the input of another Step by a channel. The flow above can be modeled as a graph of Step’s connected by channels.

The Indexer SDK’s architecture simplifies writing custom processors in several ways:

1. You can reuse Step implementations across processors which reduces duplication of common data extraction logic.
2. The SDK collects basic performance metrics, like the number of transactions processed, for each Step, which enables observability into subcomponents of the processor.
3. Since each Step is independent, you can safely customize parts of the processor without breaking the other pieces. For example, you can add additional Step’s to pre/post-process data or batch data writes. Each Step can also be tested in isolation from the rest of the processor.