How does Ponder store data?

Internally, Ponder uses a SQL database. This database stores two distinct types of data:

1. Blockchain data (blocks, transactions, and event logs)

• Persistent across deploys and hot reloads
• Populated by fetching data from an Etheruem RPC (Infura, Alchemy, a local Anvil node)
• Managed internally by the Ponder indexing engine

2. Entity data (served by the GraphQL server)

• Ephemeral, does not persist across deploys or hot reloads
• Populated by event handler functions via context.entities objects
• Managed by the user

In development, Ponder uses SQLite because it offers excellent performance and a streamlined developer experience.

In production, Ponder uses Postgres. SQLite works for small projects in production, but struggles with larger projects due to performance and reliability issues with persistent disks.