Struct SubscriberConfig 

pub struct SubscriberConfig {

    pub preconfirmations: PreconfirmationMode,
    pub canonical_head_poll_interval: Duration,
    pub flashblock_poll_interval: Duration,
    pub max_consecutive_flashblock_poll_failures: usize,
    pub max_pending_transaction_receipts_per_tick: usize,
    pub max_flashblock_rpc_requests_per_second: usize,
    pub hydrate_pending_transactions: bool,
    pub verify_log_block_context: bool,
    pub max_batch_size: usize,
    pub max_log_addresses_per_subscription: usize,
    pub max_pending_records: usize,
    pub max_pending_backfills: usize,
    pub max_backfill_log_bytes: usize,
    pub max_reconcile_requests_in_flight: usize,
    pub reconnect: SubscriberReconnectConfig,
}

Subscriber configuration.

Fields

preconfirmations: PreconfirmationMode

Flashblocks delivery policy. Provider support itself is configured by the transport’s single flashblocks endpoint flag.

canonical_head_poll_interval: Duration

Cadence for certifying sealed canonical heads while connected to a Flashblocks endpoint whose newHeads stream may contain partial heads.

flashblock_poll_interval: Duration

Optimism pending-state sampling cadence.

Base uses native newFlashblocks plus pendingLogs. Optimism providers currently expose the interoperable Flashblocks surface through pending RPC reads, so one generation-pinned sampler reads the cumulative pending block, its exact hash-addressed parent, filtered pending-block logs, and bounded exact transaction receipts.

max_consecutive_flashblock_poll_failures: usize

Consecutive pending-state request failure allowance.

A successful sampling tick resets this counter. Semantic integrity failures, such as non-monotonic transaction membership or malformed logs, are never retried through this allowance.

max_pending_transaction_receipts_per_tick: usize

Maximum pending receipts per sampling tick.

Receipts are requested by exact transaction hash in one JSON-RPC batch, because separate eth_getBlockReceipts("pending") responses can refer to a different cumulative Flashblock. The rolling total-method budget may impose a lower effective per-tick limit; with the defaults and one log filter, at most seven receipts are requested per tick.

max_flashblock_rpc_requests_per_second: usize

Pending-state RPC method budget per rolling one-second window.

The sampler reserves capacity for the pending-block, exact-parent, and filtered-log methods implied by its cadence and filter plan, plus the exact-parent canonical-head poll when block interests require it. Exact receipt hydration uses only an evenly apportioned remainder. Request timestamps enforce the ceiling across actual ticks, including delayed ticks. The default leaves headroom below common paid-provider limits of 50 requests per second.

hydrate_pending_transactions: bool

Hydrate pending transaction hashes into full bodies when possible.

verify_log_block_context: bool

Verify each canonical log’s block identity through RPC and enrich its context with the exact parent hash before delivery.

Enable this when a strict coordinator (such as a hybrid historical/live source) must prove canonical ancestry from log-only pubsub events. Verification is cached per block, so the provider is queried at most once for each distinct canonical block retained in the dedupe window. For high-volume pubsub filters, configure AlloySubscriber::with_log_verification_provider with a separate HTTP provider so verification responses cannot be starved by notifications.

max_batch_size: usize

Maximum records to emit per batch.

max_log_addresses_per_subscription: usize

Maximum distinct contract addresses placed in one provider-side log subscription. Compatible logical owner filters are fanned into address supersets up to this limit; exact owner routing still happens locally.

max_pending_records: usize

Maximum records retained across the delivery queue and hidden transaction-aware reconcile buffer. Exceeding it fails the subscriber closed until a full interest reset, because dropping an event would create an unknowable continuity gap.

max_pending_backfills: usize

Maximum lazy owner-backfill requests retained at once.

max_backfill_log_bytes: usize

Maximum approximate encoded bytes accepted from one historical log response (fixed log identity fields, topics, and data).

max_reconcile_requests_in_flight: usize

Maximum provider log requests concurrently in flight during bulk owner reconciliation.

reconnect: SubscriberReconnectConfig

Reconnect policy for WebSocket/pubsub streams.

Trait Implementations

impl Clone for SubscriberConfig

fn clone(&self) -> SubscriberConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for SubscriberConfig

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

Formats the value using the given formatter.

impl Default for SubscriberConfig

fn default() -> Self

Returns the “default value” for a type.

impl Eq for SubscriberConfig

impl PartialEq for SubscriberConfig

fn eq(&self, other: &SubscriberConfig) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for SubscriberConfig