Struct CacheResidencyReport 

pub struct CacheResidencyReport {

    pub logical_cached_tokens: u64,
    pub key_value_blocks: u64,
    pub compressed_latent_blocks: u64,
    pub device_blocks: u64,
    pub host_blocks: u64,
    pub disk_blocks: u64,
    pub current_device_bytes: u64,
    pub peak_device_bytes: u64,
    pub current_host_bytes: u64,
    pub peak_host_bytes: u64,
    pub current_disk_bytes: u64,
    pub peak_disk_bytes: u64,
    pub in_flight_write_blocks: u64,
    pub in_flight_write_bytes: u64,
    pub peak_in_flight_write_bytes: u64,
    pub in_flight_host_demotion_blocks: u64,
    pub in_flight_host_demotion_bytes: u64,
    pub peak_in_flight_host_demotion_bytes: u64,
    pub mutable_tail_bytes: u64,
    pub protected_recent_blocks: u64,
    pub protected_prefix_blocks: u64,
    pub per_layer: Vec<CacheLayerResidencyReport>,
    pub per_layer_overflow_layers: u64,
    pub per_layer_overflow: CacheLayerResidencyStats,
    pub host_promotions: u64,
    pub disk_promotions: u64,
    pub host_demotions: u64,
    pub disk_demotions: u64,
    pub transfer_bytes: u64,
    pub transfer_wait: Duration,
    pub evictions: u64,
    pub discarded_sliding_blocks: u64,
    pub block_seals: u64,
    pub tail_allocations: u64,
    pub demand_hits: u64,
    pub demand_misses: u64,
    pub in_flight_waits: u64,
    pub queue_capacity: usize,
    pub queue_peak_occupancy: usize,
    pub queue_backpressure: u64,
    pub cancellations: u64,
    pub failures: u64,
    pub prefill_full_attention_blocks: u64,
    pub prefill_full_attention_bytes: u64,
    pub decode_full_attention_blocks: u64,
    pub decode_full_attention_bytes: u64,
    pub attention_scratch_peak_bytes: u64,
    pub prompt_cache_saves: u64,
    pub prompt_cache_loads: u64,
    pub prompt_cache_bytes: u64,
    pub imported_buffered_shards: u64,
    pub process_rss_bytes: Option<u64>,
    pub process_minor_page_faults: Option<u64>,
    pub process_major_page_faults: Option<u64>,
}

Aggregated logical device/disk and physical host residency observations.

Fields

logical_cached_tokens: u64

Absolute token count represented by the longest layer.

key_value_blocks: u64

Sealed key/value blocks.

compressed_latent_blocks: u64

Sealed compressed-latent/rotary blocks.

device_blocks: u64

Blocks cataloged on the execution device.

host_blocks: u64

Blocks cataloged in host memory.

disk_blocks: u64

Blocks cataloged on disk.

current_device_bytes: u64

Current logical device bytes, including mutable tails.

peak_device_bytes: u64

Peak successfully admitted logical device bytes.

current_host_bytes: u64

Current physical host allocation capacity.

peak_host_bytes: u64

Peak successfully admitted physical host allocation capacity.

current_disk_bytes: u64

Current logical disk bytes.

peak_disk_bytes: u64

Peak successfully admitted logical disk bytes.

in_flight_write_blocks: u64

Blocks whose host buffers are owned by background disk writes.

in_flight_write_bytes: u64

Physical host capacity owned by disk writes.

peak_in_flight_write_bytes: u64

Peak physical host capacity owned by background disk writes.

in_flight_host_demotion_blocks: u64

Blocks retaining both device and host allocations during demotion.

in_flight_host_demotion_bytes: u64

Physical host capacity charged during device demotion.

peak_in_flight_host_demotion_bytes: u64

Peak physical host capacity charged during device demotion.

mutable_tail_bytes: u64

Current bytes in mutable tails.

protected_recent_blocks: u64

Recent blocks protected from device demotion.

protected_prefix_blocks: u64

Prefix or sink blocks protected for attention.

per_layer: Vec<CacheLayerResidencyReport>

Current per-layer rows, sorted and bounded.

per_layer_overflow_layers: u64

Number of active layers folded into the overflow row.

per_layer_overflow: CacheLayerResidencyStats

Exact aggregate of omitted and unidentifiable layers.

host_promotions: u64

Host-to-device promotions.

disk_promotions: u64

Disk-to-device promotions.

host_demotions: u64

Device-to-host demotions.

disk_demotions: u64

Resident-to-disk demotions.

transfer_bytes: u64

Logical bytes copied by promotions and demotions.

transfer_wait: Duration

Host time spent at transfer ownership boundaries.

evictions: u64

Blocks evicted after configured tiers were exhausted.

discarded_sliding_blocks: u64

Sliding-window blocks discarded as invisible.

block_seals: u64

Completed block seals.

tail_allocations: u64

Mutable tail allocations.

demand_hits: u64

Demand hits.

demand_misses: u64

Demand misses.

in_flight_waits: u64

Requests that joined an existing transfer.

queue_capacity: usize

Effective bounded disk request queue capacity.

queue_peak_occupancy: usize

Peak observed queue occupancy.

queue_backpressure: u64

Requests delayed by queue capacity.

cancellations: u64

Requests canceled by reset or truncation.

failures: u64

Cache transfer or persistence failures.

prefill_full_attention_blocks: u64

Blocks scanned by full attention during prefill.

prefill_full_attention_bytes: u64

Logical bytes scanned by full attention during prefill.

decode_full_attention_blocks: u64

Blocks scanned by full attention during decode.

decode_full_attention_bytes: u64

Logical bytes scanned by full attention during decode.

attention_scratch_peak_bytes: u64

Peak logical scratch bytes used by attention.

prompt_cache_saves: u64

Successful prompt-cache saves.

prompt_cache_loads: u64

Successful prompt-cache loads.

prompt_cache_bytes: u64

Logical bytes written or cataloged for prompt caches.

imported_buffered_shards: u64

Imported persistent shard count.

process_rss_bytes: Option<u64>

Optional peak process resident-set size.

process_minor_page_faults: Option<u64>

Optional cumulative minor page faults.

process_major_page_faults: Option<u64>

Optional cumulative major page faults.

Trait Implementations

impl Clone for CacheResidencyReport

fn clone(&self) -> CacheResidencyReport

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CacheResidencyReport

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

Formats the value using the given formatter.

impl Default for CacheResidencyReport

fn default() -> CacheResidencyReport

Returns the “default value” for a type.

impl Eq for CacheResidencyReport

impl PartialEq for CacheResidencyReport

fn eq(&self, other: &CacheResidencyReport) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for CacheResidencyReport