Struct InvalidationConfig 

pub struct InvalidationConfig {
    pub debounce_ms: u64,
    pub max_debounce_ms: u64,
    pub coalesce_by_table: bool,
    pub max_buffer_size: usize,
}

Configuration for the invalidation engine.

The invalidation engine uses a debounce algorithm to batch rapid changes into single re-executions. This prevents “thundering herd” scenarios where a batch insert of 1000 rows would trigger 1000 subscription refreshes.

The algorithm works as follows:

1. When a change arrives, record the subscription as pending
2. Wait for debounce_ms of silence (no new changes to that subscription)
3. If max_debounce_ms passes since the first change, flush anyway
4. If buffer exceeds max_buffer_size, flush immediately (memory protection)

This balances latency (users want updates fast) against efficiency (batching reduces database load). Default values target 50ms debounce with 200ms max wait, meaning updates arrive within 200ms worst-case.

Fields

debounce_ms: u64

Debounce window in milliseconds. After a change, wait this long for more changes before invalidating.

max_debounce_ms: u64

Maximum debounce wait in milliseconds. Even if changes keep arriving, invalidate after this duration.

coalesce_by_table: bool

Whether to coalesce changes by table. When true, multiple changes to the same table become a single invalidation.

max_buffer_size: usize

Maximum changes to buffer before forcing flush. Prevents unbounded memory growth during high-throughput periods.

Trait Implementations

impl Clone for InvalidationConfig

fn clone(&self) -> InvalidationConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for InvalidationConfig

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

Formats the value using the given formatter.

impl Default for InvalidationConfig

fn default() -> Self

Returns the “default value” for a type.