Struct ResourceUsage 

pub struct ResourceUsage {
    pub memory_usage_mb: usize,
    pub total_memory_mb: usize,
    pub process_memory_mb: Option<usize>,
    pub cpu_usage_percent: f64,
    pub cpu_usage_percent_valid: bool,
    pub gpu_usage_percent: Option<f64>,
    pub network_io_mbps: Option<f64>,
    pub disk_io_mbps: Option<f64>,
    pub active_threads: usize,
    pub timestamp: Instant,
}

Current resource usage information

Fields

memory_usage_mb: usize

Memory usage in MB

total_memory_mb: usize

Total system memory in MB (R2 fix). Populated from the real system total at collection time so memory_usage_mb can be turned into an honest percentage instead of assuming a fixed 1024 MB (1 GB) total, which pinned memory_percent/memory_utilization above 100% (and therefore alert severity at Emergency) on essentially every real machine. 0 (the Default value) means “unknown”; callers must treat that as “cannot compute a percentage”, not as “0 MB total”.

process_memory_mb: Option<usize>

Memory this process is using, in MB, or None when the process counters are unavailable.

memory_usage_mb/total_memory_mb are system-wide and only meaningful as a ratio (which is what the memory alert thresholds compare). The allocation budget, by contrast, is a per-process figure, so comparing system-wide usage against it reports several hundred percent utilization on any real machine — the same class of error as R2, one level up. That comparison was inert only because start_monitoring was never called and current_usage stayed all-zero; every budget consumer now reads this field instead.

cpu_usage_percent: f64

CPU usage percentage (0-100).

R1: this used to be the hardcoded literal 50.0. It is now the real system-wide CPU usage read from a persistent sysinfo::System, because CPU usage is a delta between two refreshes: a freshly built System always reports 0. Until two refreshes at least sysinfo::MINIMUM_CPU_UPDATE_INTERVAL apart have happened this field holds 0.0 and cpu_usage_percent_valid is false; consumers must check the flag rather than treating the placeholder as a measurement.

cpu_usage_percent_valid: bool

Whether cpu_usage_percent holds a real measurement.

gpu_usage_percent: Option<f64>

GPU usage percentage (0-100) if applicable. There is no pure-Rust, vendor-neutral way to read this, so it stays None unless an external probe supplies it.

network_io_mbps: Option<f64>

Network I/O rate in MB/s, or None before two interface refreshes have produced a usable byte delta (R1: this used to be the literal 1.0).

disk_io_mbps: Option<f64>

Disk I/O rate in MB/s for this process, or None before two process refreshes have produced a usable byte delta (R1: this used to be the literal 5.0).

active_threads: usize

Number of active threads

timestamp: Instant

Timestamp of measurement

Implementations

impl ResourceUsage

pub fn process_memory(&self) -> Option<usize>

Memory this process is using, in MB, when it could be measured.

pub fn cpu_usage(&self) -> Option<f64>

Real CPU usage, or None when no valid measurement has been taken yet.

pub fn memory_usage_percent(&self) -> Option<f64>

Real memory-usage percentage (R2), used / total * 100. Returns None when total_memory_mb is unknown (0) rather than fabricating a value against a wrong assumed total.

Trait Implementations

impl Clone for ResourceUsage

fn clone(&self) -> ResourceUsage

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ResourceUsage

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

Formats the value using the given formatter.

impl Default for ResourceUsage

fn default() -> Self

Returns the “default value” for a type.

impl Serialize for ResourceUsage

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.