Struct RpcSelector

pub struct RpcSelector { /* private fields */ }

Manages selection of RPC endpoints based on configuration.

Implementations

impl RpcSelector

pub fn new( configs: Vec<RpcConfig>, failure_threshold: u32, pause_duration_secs: u64, failure_expiration_secs: u64, ) -> Result<Self, RpcSelectorError>

Creates a new RpcSelector instance.

Arguments

• configs - RPC configurations
• failure_threshold - Number of consecutive failures before pausing a provider. Defaults to [DEFAULT_PROVIDER_FAILURE_THRESHOLD] if not provided via env var PROVIDER_FAILURE_THRESHOLD.
• pause_duration_secs - Duration in seconds to pause a provider after reaching failure threshold. Defaults to [DEFAULT_PROVIDER_PAUSE_DURATION_SECS] if not provided via env var PROVIDER_PAUSE_DURATION_SECS.
• failure_expiration_secs - Duration in seconds after which failures are considered stale and reset. Defaults to [DEFAULT_PROVIDER_FAILURE_EXPIRATION_SECS] (60 seconds).

Returns

• Result<Self> - A new selector instance or an error

Note

These values are typically loaded from ServerConfig::from_env() which reads from environment variables:

• PROVIDER_FAILURE_THRESHOLD (default: 3, legacy: RPC_FAILURE_THRESHOLD)
• PROVIDER_PAUSE_DURATION_SECS (default: 60, legacy: RPC_PAUSE_DURATION_SECS)

pub fn new_with_defaults( configs: Vec<RpcConfig>, ) -> Result<Self, RpcSelectorError>

Creates a new RpcSelector instance with default failure threshold and pause duration.

This is a convenience method primarily for testing. In production code, use new() with values from ServerConfig::from_env().

Arguments

• configs - RPC configurations

Returns

• Result<Self> - A new selector instance or an error

pub fn provider_count(&self) -> usize

Gets the number of available providers

Returns

• usize - The number of providers in the selector

pub fn available_provider_count(&self) -> usize

Gets the number of available (non-paused) providers

Returns

• usize - The number of non-paused providers

pub fn get_configs(&self) -> Vec<RpcConfig>

Gets the current RPC configurations.

Returns

• Vec<RpcConfig> - The current configurations

pub fn mark_current_as_failed(&self)

Marks the current endpoint as failed and forces selection of a different endpoint.

This method is used when a provider consistently fails, and we want to try a different one. It adds the current provider to the failed providers set and will avoid selecting it again.

pub fn get_current_url(&self) -> Result<String, RpcSelectorError>

Gets the URL of the currently selected RPC endpoint.

Returns

• Result<String, RpcSelectorError> - The URL of the current provider, or an error

pub fn get_next_url( &self, excluded_urls: &HashSet<String>, ) -> Result<String, RpcSelectorError>

Gets the URL of the next RPC endpoint, excluding providers that have already been tried.

Arguments

• excluded_urls - URLs of providers that have already been tried in the current failover cycle

Returns

• Result<String, RpcSelectorError> - The URL of the next provider, or an error

pub fn get_client<T>( &self, initializer: impl Fn(&str) -> Result<T>, excluded_urls: &HashSet<String>, ) -> Result<T, RpcSelectorError>

Gets a client for the selected RPC endpoint.

Arguments

• initializer - A function that takes a URL string and returns a Result<T>
• excluded_urls - URLs of providers that have already been tried in the current failover cycle

Returns

• Result<T> - The client instance or an error

Trait Implementations

impl Clone for RpcSelector

fn clone(&self) -> Self

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for RpcSelector

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.