Struct RetryCauseTelemetry 

pub struct RetryCauseTelemetry { /* private fields */ }

Counts retried requests by what caused the retry.

Attach this to StandardHttpRetryLifecycle to count retries as they’re decided, at the point where the response or the error that triggered them is still available.

Metrics

• network_http_requests_retry_causes_total: the total number of retried requests, tagged with the domain given at construction and with a bounded description of the failure:
  • cause: one of http_status, http2, tls, timeout, connection, client, or other.
  • frame and initiator: for cause:http2 only. frame is the HTTP/2 frame the error came from, and initiator is the side of the connection that sent it, where local means this client ended the request rather than the remote endpoint.
  • reason: for cause:http2, the HTTP/2 reason code by name; for cause:timeout, the stage that timed out; for cause:connection, the kind of transport failure. Absent when the failure carries no reason we recognize.

Retries caused by response status codes aren’t broken down by code here, since network_http_requests_errors_total already counts every non-success response by status code.

Implementations

impl RetryCauseTelemetry

pub fn from_builder(builder: &MetricsBuilder, domain: &str) -> Self

Creates retry cause telemetry for requests sent to domain.

The domain is used as-is for the domain tag, so pass the same value that the rest of a client’s telemetry uses: the scheme, host, and port of the remote endpoint.

Trait Implementations

impl Clone for RetryCauseTelemetry

fn clone(&self) -> RetryCauseTelemetry

Returns a duplicate of the value.

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

Performs copy-assignment from source.