Performance & Scalability

When designing our indexers, we prioritized high-performance historical data backfilling to ensure developers can iterate quickly. Instead of waiting weeks for an indexer to catch up with real-time data, our approach accelerates the process, enabling faster development cycles.

Two Execution Modes: Realtime & Historical

Indexers operate in two distinct modes:

• Realtime Mode – Functions like a traditional sequential indexer, processing new blockchain data as it arrives.
• Historical Mode – Enables parallelized indexing of past data, significantly reducing the time required for backfilling.

This dual-mode execution is made possible by a state system based on change logs, allowing multiple threads to process data without compromising consistency.

State Management & Chain Reorganizations

By leveraging change logs for state management, our system automatically handles blockchain reorganizations in real-time, eliminating the need for manual intervention by developers. This ensures accurate and reliable indexing, even in dynamic blockchain environments.

Versioned Outputs & Stream Replay

To further enhance reliability, we store versioned outputs of indexers in cloud storage. This allows developers to replay streams seamlessly—provided the indexer code and configuration remain unchanged—without the need for redundant reprocessing.