Struct Gradient2 

pub struct Gradient2 { /* private fields */ }

Netflix Gradient2–inspired adaptive concurrency limit strategy.

Instead of estimating absolute queue depth (like Vegas), Gradient2 uses the gradient (ratio) of long-term RTT to short-term RTT to detect queueing. A configurable tolerance allows small RTT increases without triggering limit reduction, making it more robust to natural latency variance.

The algorithm maintains two RTT estimates:

• long RTT: exponentially smoothed baseline (adapts slowly)
• short RTT: most recent observed RTT (reacts immediately)

Each update computes:

gradient  = clamp(tolerance × long_rtt / short_rtt, 0.5, 1.0)
new_limit = gradient × current_limit + queue_size
limit     = smooth(current_limit, new_limit)

Implementations

impl Gradient2

pub fn builder() -> Gradient2Builder

Returns a Gradient2Builder for configuring a new Gradient2 instance.

Trait Implementations

impl Algorithm for Gradient2

fn max_concurrency(&self) -> usize

Returns the maximum number of concurrent requests the algorithm currently allows.

fn update( &mut self, rtt: Duration, num_inflight: usize, _is_error: bool, is_canceled: bool, )

Observes the outcome of a request and updates the algorithm’s state accordingly.

impl Clone for Gradient2

fn clone(&self) -> Gradient2

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Gradient2

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

Formats the value using the given formatter.

impl Default for Gradient2

fn default() -> Self

Returns the “default value” for a type.