Struct VxKernel 

pub struct VxKernel { /* private fields */ }

The VX Kernel — manages token lifecycle and dispatches work

The kernel operates in a loop:

1. Exhaust ready queue (no await — pure CPU work)
2. Check completion (all done? deadlock?)
3. Park and wait (single await point: select! on lanes)

Implementations

impl VxKernel

pub fn new(service: impl Into<String>) -> Self

Create a new kernel for a service

pub fn service(&self) -> &str

Get the service name

pub fn metrics(&self) -> &Arc<KernelMetrics>

Get metrics reference

pub fn is_accepting(&self) -> bool

Check if kernel is accepting requests

pub fn pause(&self)

Pause accepting (backpressure)

pub fn resume(&self)

Resume accepting

pub fn enqueue( &self, request_id: u64, service: String, payload: Vec<u8>, ) -> bool

Enqueue a new request token

pub fn dequeue_ready(&self) -> Option<u64>

Dequeue next ready token (non-blocking)

pub fn complete(&self, request_id: u64)

Mark a token as completed

pub fn fail(&self, request_id: u64)

Mark a token as failed

pub fn cancel(&self, request_id: u64)

Cancel a token (e.g., client disconnect)

pub fn token_state(&self, request_id: u64) -> Option<TokenState>

Get token state

pub fn token_payload(&self, request_id: u64) -> Option<Vec<u8>>

Get token payload

pub fn in_flight(&self) -> usize

Get current in-flight count

pub fn ready_count(&self) -> usize

Get ready queue length

pub fn handle_control(&self, signal: ControlSignal)

Process a control signal

pub fn cleanup(&self, max_age_nanos: u64) -> usize

Clean up completed/failed/cancelled tokens older than max_age_nanos