Struct FailureTracker 

pub struct FailureTracker { /* private fields */ }

Tracks persistently-failing items so they don’t permanently consume batch slots. After [FAILURE_THRESHOLD] consecutive failures for the same key, the item is skipped for [BACKOFF_CYCLES] scan cycles.

Implementations

impl FailureTracker

pub fn new() -> Self

pub fn advance_cycle(&self)

Call once per full scan cycle (e.g., when partition == 0).

pub fn should_skip(&self, key: &str) -> bool

Returns true if this item should be skipped (in backoff). Also resets the entry when backoff expires, giving it another chance.

pub fn record_failure(&self, key: &str, scanner_name: &str)

Record a failure. After [FAILURE_THRESHOLD] consecutive failures, logs an error and puts the item into backoff.

pub fn record_success(&self, key: &str)

Record a success — clears any tracked failure state.

Trait Implementations

impl Default for FailureTracker

fn default() -> FailureTracker

Returns the “default value” for a type.