Struct TraverseConfig 

pub struct TraverseConfig { /* private fields */ }

Configuration for an indexer.

Implementations

impl TraverseConfig

pub fn new_single(endpoint: Endpoint, start_height: AbsoluteBlockHeight) -> Self

A configuration with a single endpoint starting at the given block height.

pub fn new( endpoints: Vec<Endpoint>, start_height: AbsoluteBlockHeight, ) -> Option<Self>

A configuration with a given list of endpoints and starting height. Returns None if the list of endpoints is empty.

pub fn set_max_behind(self, max_behind: Duration) -> Self

Set the maximum number of time the last finalized block of the node can be behind before it is deemed too far behind, and another node is tried.

The default value is 60 seconds.

pub fn set_wait_after_failure(self, wait_after_fail: Duration) -> Self

After each failure the indexer will pause a bit before trying another node to avoid overloading the node. Defaults to 1 second.

pub fn push_endpoint(self, endpoint: Endpoint) -> Self

Add an additional endpoint to the list of endpoints the indexer will use. This is added to the end of the list, so the endpoint is only tried in case of failure of previous queries.

pub fn set_max_parallel(self, max_parallel: usize) -> Self

Set the maximum number of blocks that will be queried in parallel, if they are available. Defaults to 4 if not set explicitly.

pub async fn traverse<I: Indexer>( self, indexer: I, sender: Sender<I::Data>, ) -> TraverseResult<()>

Traverse the chain according to the supplied configuration, invoking on_finalized for each finalized block.

Multiple on_finalized calls might be executed concurrently, but their responses will be written to the provided tokio::sync::mpsc::Sender in the increasing order of block height, with no gaps.

If a query fails, either due to timeout or node not being available the indexer will attempt the next endpoint it is configured with.

For robust indexing it is crucial that the supplied Endpoints are configured with timeouts so that the indexer may make progress.

The traverse method will return either when signalled so by the on_failure method or when the receiver part of the supplied tokio::sync::mpsc::Sender is closed. Typically this method should run in a task spawned via tokio::spawn.