pub fn flush_telemetry(
timeout_seconds: Option<f64>,
) -> Result<(), TelemetryError>Expand description
Force-flush installed providers without tearing them down.
The drain half of shutdown_telemetry: every provider we installed is
force-flushed under the bounded-shutdown deadline
(PROVIDE_EXPORTER_LOGS_SHUTDOWN_TIMEOUT_SECONDS) and stays installed and
usable. Use it where records must be out before control returns — a request
boundary, a checkpoint, a serverless freeze — rather than shutting telemetry
down and paying to set it up again.
Returns Ok(()) when every signal drained within the deadline (including
when nothing is installed) and Err when any was abandoned, so a caller
flushing to be sure its records are out learns when they are not.
timeout_seconds overrides the configured deadline for this call; None
uses the configured one. Matches Python’s flush_telemetry(timeout_seconds),
TypeScript’s flushTelemetry(timeoutMs) and Go’s context deadline.