Struct ServiceRuntime 

pub struct ServiceRuntime {
    pub metrics: MetricsManager,
    pub dfe: Arc<DfeMetrics>,
    pub memory_guard: Arc<MemoryGuard>,
    pub shutdown: CancellationToken,
    pub context: &'static RuntimeContext,
    pub worker_pool: Option<Arc<AdaptiveWorkerPool>>,
    pub batch_engine: Option<Arc<BatchEngine>>,
    pub scaling: Option<Arc<ScalingPressure>>,
}

Available on crate feature cli only.

Pre-built service infrastructure. Created by run_app() before run_service().

Apps receive this fully wired – they just use the fields. No boilerplate needed.

On bare metal, K8s-specific features (pre-stop delay, pod metadata in logs) are automatically disabled. On K8s, they’re automatically enabled.

Fields

metrics: MetricsManager

Metrics manager – already started, serving endpoints. Use for registering app-specific metrics and metric groups.

dfe: Arc<DfeMetrics>

Platform DFE metrics (dfe_* counters/gauges). Already registered.

memory_guard: Arc<MemoryGuard>

Available on crate feature memory only.

Cgroup-aware memory guard. Tracks memory usage for backpressure. Auto-detected from env prefix + cgroup limits.

shutdown: CancellationToken

Shutdown token. Cancelled on SIGTERM/SIGINT (with K8s pre-stop delay). Clone and pass to your pipeline loops.

context: &'static RuntimeContext

Runtime context – K8s/Docker/BareMetal metadata (pod_name, namespace, etc.).

worker_pool: Option<Arc<AdaptiveWorkerPool>>

Available on crate feature worker-pool only.

Adaptive worker pool for parallel batch processing (worker feature). None if the worker feature is not enabled or config fails.

batch_engine: Option<Arc<BatchEngine>>

Available on crate feature worker-batch only.

Batch processing engine with SIMD parsing and pre-route filtering (worker-batch feature). None if worker-batch is not enabled or worker pool creation failed.

scaling: Option<Arc<ScalingPressure>>

Available on crate feature scaling only.

Scaling pressure calculator for KEDA autoscaling (scaling feature). None if the scaling feature is not enabled.

Implementations

impl ServiceRuntime

pub fn set_readiness_check<F: Fn() -> bool + Send + Sync + 'static>( &mut self, check: F, )

Set the readiness check callback.

Each app defines its own readiness criteria. Call this in run_service() once you know what “ready” means for your app.

pub fn batch_engine(&self) -> Option<&Arc<BatchEngine>>

Available on crate feature worker-batch only.

Return the batch processing engine, if the worker-batch feature is enabled and the worker pool was successfully created.