Chainhooks use and discussion

[mefrem]

As an early user, I want to share my notes and get any feedback about using Chainhooks.

Chainhooks is the answer to the question, "How do I react to events occurring on the Stacks or Bitcoin chain?"

• When you create an app, you create a smart contract that updates the state of the blockchain based on transactions made by users. If I build a smart contract with a token A that I want to swap for token B, I will be sending a transaction where a method of the smart contract will be called, and this method will be taking some tokens from my balance and giving me some other token(s). The state is updated.
• If I want to react to these state changes, I need to come up with a solution that observes the blockchain, and poll the state of the chain.
• Polling is a bit hard to implement, plus blockchains can fork.
• The way the chainhooks layer works, is as a developer I am going to register a predicate that will be evaluated every time a new block is mined. When this predicate appears to be true, then I will be given an alert. Then based on that, I can take some further action.

Here's a list of predicates under the chainhook-node folder

Predicates can be categorized as any action that takes place on the blockchain, like for Stacks

• txid for a specific transaction
• ft-events like minting and burning fungible tokens
• nft-events like transfer and burning
• stx-event like minting or locked for PoX
• print-events for contracts that emit print events, or
• contract-deploy with either a specific deployer or a trait included.

And/or for monitoring the Bitcoin chain

• tx-id
• outputs - OP_RETURN output ending with a set of characters, p2pkh payments,
  Chainhooks available as a self-hosted solution, as well as Hiro hosted solution in the future.

Some more notes:

• Chainhooks is not a websocket Stacks API. The websocket in the API is unfortunately shadowing fork events, which means if your transaction is mined, you will be getting a notification. But if there is a fork, then there would be some issues with consistency
• In chainhooks, you receive a payload with two loads, one is "APPLY" with a set of transactions to apply, and "ROLLBACK" that provides tx to rollback if there is a fork.
• Stacks blockchain API was meant to index a certain set of data, whereas this Chainhooks layer is designed to help you index your own data.
• Right now we have a stateless chainhooks, which means you can only evaluate predicates on blocks being streamed, which makes it available for only new protocols.
• @lgalabru is wrapping up stateful version, which gives ability to scan history of chain, and is coming sometime in the next quarter