Struct QueryDiagnostics 

#[non_exhaustive]
pub struct QueryDiagnostics {

    pub snapshot_generation: u64,
    pub indexed_through_seq: LogSeq,
    pub plan: Vec<SegmentPlan>,
    pub approximate: bool,
    pub exact_rescore: bool,
    pub scan_rescore: Option<ScanRescoreCounters>,
    pub materialization: Option<MaterializationCounters>,
    pub fusion: Option<FusionReport>,
    pub hybrid: Option<HybridReport>,
    pub hybrid_tier_resolution: Option<HybridTierResolution>,
    pub requested_k: usize,
    pub returned: usize,
    pub budget_exhausted: bool,
    pub counters: QueryCounters,
    pub embedding_epoch: Option<EpochId>,
    pub tokenizer_epoch: Option<TokenizerEpoch>,
    pub observed_qos: ObservedQos,
    pub elapsed: Duration,
    pub timings: Option<QueryTimings>,
}

Version 1 unconditional report of what one store query actually did.

Fields (Non-exhaustive)

Non-exhaustive structs could have additional fields added in future. Therefore, non-exhaustive structs cannot be constructed in external crates using the traditional Struct { .. } syntax; cannot be matched against without a wildcard ..; and struct update syntax will not work.

snapshot_generation: u64

Version 1 generation pinned for the complete query.

indexed_through_seq: LogSeq

Highest WAL sequence represented by the pinned active and sealed state.

plan: Vec<SegmentPlan>

Version 1 per-segment branches in execution order.

approximate: bool

True when any candidate membership came from graph traversal or another non-exhaustive path.

exact_rescore: bool

True when every returned score was computed from full-precision rows.

scan_rescore: Option<ScanRescoreCounters>

Work from explicit quantized candidate scanning and selected-row rescoring.

materialization: Option<MaterializationCounters>

Scoped result hydration work; absent for hit-only core queries.

fusion: Option<FusionReport>

Version 1 fusion report; absent when no hybrid fusion leg ran.

hybrid: Option<HybridReport>

Bounded-producer facts; absent when no hybrid fusion leg ran.

hybrid_tier_resolution: Option<HybridTierResolution>

Resolution applied when a hybrid caller supplied no tier preference. Explicit caller tiers leave this absent.

requested_k: usize

Requested result count.

returned: usize

Result count actually returned.

budget_exhausted: bool

True when an execution budget fired, including an exact fallback.

counters: QueryCounters

Version 1 exact counter composition.

embedding_epoch: Option<EpochId>

Embedding identity that interpreted a vector leg.

tokenizer_epoch: Option<TokenizerEpoch>

Tokenizer identity that interpreted a lexical leg.

observed_qos: ObservedQos

Caller-thread QoS observed without requesting a scheduling change.

elapsed: Duration

Wall time spent in the admitted query path.

timings: Option<QueryTimings>

Optional stage clocks; absent unless built with query-timing.

Trait Implementations

impl Clone for QueryDiagnostics

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for QueryDiagnostics

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

Formats the value using the given formatter.

impl PartialEq for QueryDiagnostics

Compares semantic query behavior while intentionally excluding elapsed, because wall-clock timing varies between otherwise identical searches. The exhaustive destructuring makes every future field an explicit equality choice.

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

Equality operator ==.

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

Inequality operator !=.