Struct RetryPolicy 

pub struct RetryPolicy { /* private fields */ }

Configuration for client retry behavior.

Examples

use std::time::Duration;
use openkind_client::RetryPolicy;

// Three quick retries, honoring the server's Retry-After.
let policy = RetryPolicy::new()
    .max_retries(3)
    .backoff_initial(Duration::from_millis(100))
    .backoff_max(Duration::from_secs(2));

Implementations

impl RetryPolicy

pub fn new() -> Self

A policy with the SDK-default settings (same as RetryPolicy::default).

pub fn max_retries(self, max_retries: u32) -> Self

Maximum retries after the initial attempt; 0 disables retries.

pub fn backoff_initial(self, initial: Duration) -> Self

First backoff delay, doubled each attempt up to backoff_max. Zero disables backoff delays entirely.

pub fn backoff_max(self, max: Duration) -> Self

Upper bound on exponential backoff delays. Zero disables backoff delays entirely.

pub fn backoff_jitter(self, jitter: f64) -> Self

Fraction of each backoff delay randomly subtracted, in 0.0..=1.0. 0.0 makes delays fully deterministic (useful in tests).

pub fn retry_status(self, status: u16) -> Self

Retry these status codes (in addition to 5xx while retry_server_errors is set). The default retries 408 and 429.

pub fn with_retry_statuses( self, statuses: impl IntoIterator<Item = u16>, ) -> Self

Replace the retried-status list wholesale, mirroring the Python SDK’s http_statuses=[...] semantics: only these codes (plus 5xx while retry_server_errors is set) are retried.

RetryPolicy::new().with_retry_statuses([409]).retry_server_errors(false) retries 409 and nothing else.

pub fn retry_predicate(self, predicate: Option<RetryPredicate>) -> Self

Opt extra errors into retrying, mirroring the Python SDK’s predicate/exceptions options: Some(predicate) retries any error for which the predicate returns true, even non-transient statuses like 404. None (the default) uses only the built-in rules.

pub fn retry_server_errors(self, retry: bool) -> Self

Whether any 5xx status (including 529 Overloaded) is retried. Default true.

pub fn respect_retry_after(self, respect: bool) -> Self

Whether to honor Retry-After / retry-after-ms response headers instead of the computed backoff. Default true.

pub fn retry_connection_errors(self, retry: bool) -> Self

Whether connection failures are retried. Default true.

pub fn retry_timeout_errors(self, retry: bool) -> Self

Whether per-attempt timeouts are retried. Default true.

pub fn total_timeout(self, timeout: Option<Duration>) -> Self

Total budget per call, including the initial attempt and all delays. A retry whose delay would reach or exceed the remaining budget is not attempted; the last error is returned instead. None disables the limit.

pub fn is_retryable(&self, error: &Error) -> bool

Whether this error should be retried under the policy.

pub fn delay_for(&self, error: &Error, retry_index: u32) -> Duration

How long to wait before the retry numbered retry_index (0-based: 0 is the wait before the first retry).

A server-requested Retry-After takes precedence over computed backoff when respect_retry_after is set.

pub fn backoff_delay(&self, retry_index: u32) -> Duration

Exponential backoff with subtractive jitter: min(backoff_max, backoff_initial * 2^retry_index) * (1 - U * jitter).

Nanosecond precision throughout, so sub-millisecond initial delays and very large caps behave like the Python SDK’s float math.

Trait Implementations

impl Clone for RetryPolicy

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for RetryPolicy

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

Formats the value using the given formatter.

impl Default for RetryPolicy

fn default() -> Self

Returns the “default value” for a type.