Struct ReconnectConfig 

pub struct ReconnectConfig {
    pub initial_backoff: Duration,
    pub max_backoff: Duration,
    pub max_attempts: usize,
}

Operator-supplied reconnect backoff settings.

The settings govern both session establishment and the run loop’s cumulative mid-run session-drop budget.

Budget reset: the cumulative drop budget resets to zero once an established session proves healthy — it served at least one task, or it stayed connected longer than max_backoff (measured monotonically from successful registration to the moment the stream ended or dropped; post-drop draining of in-flight activities never extends it). The cap is the policy’s own definition of the longest pause, so a session outliving it is demonstrably past the flapping regime, and a served task proves end-to-end health. A genuinely flapping server — no session ever serves a task or outlives max_backoff — exhausts the budget after exactly max_attempts drops.

Drains and clean closes: a server-announced drain (the wire DrainRequest frame) is an unbudgeted drop — the worker finishes in-flight work and redials after initial_backoff; the drain classification latches for the session, so even an abrupt end after the frame stays drain-class. An unannounced clean stream close remains a budgeted retryable drop: the worker redials through the same budgeted, backed-off cycle, and only a persistent unannounced clean-close loop exhausts the budget (surfacing crate::error::WorkerError::CleanCloseExhausted).

Shutdown during a drop backoff: every SDK races the backoff sleep against the shutdown signal and returns promptly, and the run outcome is aligned across the Rust, Python, and TypeScript workers: a pending drain-class or clean-close drop ends the run cleanly, while a pending error-class drop surfaces its error — a supervisor sees “this worker was mid-fault” distinctly from “this worker drained cleanly”.

Fields

initial_backoff: Duration

Initial reconnect backoff delay. Must be non-zero before reconnecting.

max_backoff: Duration

Maximum reconnect backoff delay cap. Must be non-zero before reconnecting. Doubles as the session-health threshold for the drop-budget reset described on this type.

max_attempts: usize

Maximum reconnect attempts before surfacing the last connection error.

Implementations

impl ReconnectConfig

pub const fn new( initial_backoff: Duration, max_backoff: Duration, max_attempts: usize, ) -> Self

Creates reconnect settings with every field supplied explicitly.

Trait Implementations

impl Clone for ReconnectConfig

fn clone(&self) -> ReconnectConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ReconnectConfig

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

Formats the value using the given formatter.

impl Eq for ReconnectConfig

impl PartialEq for ReconnectConfig

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl StructuralPartialEq for ReconnectConfig