Struct RuntimeMetrics 

pub struct RuntimeMetrics {

    pub registry: Registry,
    pub vm_boot_duration: Histogram,
    pub vm_boot_phase_duration: HistogramVec,
    pub vm_count: IntGaugeVec,
    pub vm_created_total: IntCounter,
    pub vm_destroyed_total: IntCounter,
    pub vm_cpu_percent: GaugeVec,
    pub vm_memory_bytes: GaugeVec,
    pub exec_total: IntCounter,
    pub exec_duration: Histogram,
    pub exec_errors_total: IntCounter,
    pub image_pull_total: IntCounter,
    pub image_pull_duration: Histogram,
    pub image_build_total: IntCounter,
    pub rootfs_cache_hits: IntCounter,
    pub rootfs_cache_misses: IntCounter,
    pub warm_pool_size: IntGauge,
    pub warm_pool_capacity: IntGauge,
    pub warm_pool_hits: IntCounter,
    pub warm_pool_misses: IntCounter,
    pub warm_pool_boots_inflight: IntGauge,
    pub warm_pool_boot_failures_total: IntCounter,
    pub warm_pool_initial_fill_duration: Histogram,
}

Pre-registered Prometheus metrics for the Box runtime.

Fields

registry: Registry

Prometheus registry holding all metrics.

vm_boot_duration: Histogram

VM boot duration in seconds.

vm_boot_phase_duration: HistogramVec

VM boot duration split by a stable lifecycle phase.

The phase label is intentionally bounded to runtime-owned values such as layout, prepare, launch, and readiness; callers must not use box IDs, image references, or other unbounded values as labels.

vm_count: IntGaugeVec

Number of VMs by state (created, ready, busy, compacting, stopped).

vm_created_total: IntCounter

Total VMs created since process start.

vm_destroyed_total: IntCounter

Total VMs destroyed since process start.

vm_cpu_percent: GaugeVec

VM CPU usage percentage (per VM, labeled by box_id).

vm_memory_bytes: GaugeVec

VM memory usage in bytes (per VM, labeled by box_id).

exec_total: IntCounter

Total exec commands executed.

exec_duration: Histogram

Exec command duration in seconds.

exec_errors_total: IntCounter

Exec commands that failed (non-zero exit or error).

image_pull_total: IntCounter

Total image pulls.

image_pull_duration: Histogram

Image pull duration in seconds.

image_build_total: IntCounter

Total image builds.

rootfs_cache_hits: IntCounter

Rootfs cache hits.

rootfs_cache_misses: IntCounter

Rootfs cache misses.

warm_pool_size: IntGauge

Current warm pool size (idle VMs).

warm_pool_capacity: IntGauge

Warm pool capacity (max_size).

warm_pool_hits: IntCounter

Total VMs allocated from warm pool.

warm_pool_misses: IntCounter

Total VMs created fresh (warm pool miss).

warm_pool_boots_inflight: IntGauge

Number of VM boots currently in flight across warm pools.

warm_pool_boot_failures_total: IntCounter

Total warm-pool VM boot failures, including failed replenishment attempts.

warm_pool_initial_fill_duration: Histogram

Time spent on the initial warm-pool fill before the pool is published.

For lazy pools this is the time until the first ready VM (the remaining idle target is filled asynchronously); eager pools report the complete configured initial fill.

Implementations

impl RuntimeMetrics

pub fn new() -> Self

Create and register all metrics with a new registry.

pub fn try_new() -> Result<Self, PrometheusError>

Try to create and register all metrics with a new registry.

pub fn with_registry(registry: Registry) -> Self

Create and register all metrics with an existing registry.

pub fn try_with_registry(registry: Registry) -> Result<Self, PrometheusError>

Try to create and register all metrics with an existing registry.

pub fn record_vm_boot_phase(&self, phase: &str, duration_secs: f64)

Record one bounded VM boot phase in seconds.

pub fn remove_vm_resource_metrics(&self, box_id: &str)

Remove resource gauges for a VM that no longer exists.

The gauges are labelled by box ID for point-in-time inspection. Keeping every historical ID in a long-lived daemon would make the registry grow without bound, so lifecycle teardown must delete both label sets.

pub fn encode(&self) -> String

Encode all metrics in Prometheus text exposition format.

Trait Implementations

impl Clone for RuntimeMetrics

fn clone(&self) -> RuntimeMetrics

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for RuntimeMetrics

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

Formats the value using the given formatter.

impl Default for RuntimeMetrics

fn default() -> Self

Returns the “default value” for a type.

impl MetricsCollector for RuntimeMetrics

fn record_vm_boot(&self, duration_secs: f64)

Record a VM boot duration in seconds.

fn inc_vm_state(&self, state: &str)

Increment the count of VMs in a given state.

fn dec_vm_state(&self, state: &str)

Decrement the count of VMs in a given state.

fn inc_vm_created(&self)

Increment total VMs created.

fn inc_vm_destroyed(&self)

Increment total VMs destroyed.

fn record_exec(&self, duration_secs: f64, success: bool)

Record an exec command completion.

fn inc_cache_hit(&self)

Record a rootfs cache hit.

fn inc_cache_miss(&self)

Record a rootfs cache miss.