Struct QueryAttemptReport 

pub struct QueryAttemptReport {

    pub schema: String,
    pub attempt_id: AttemptId,
    pub query_id: Option<AttemptQueryId>,
    pub trace_context: Option<AttemptTraceContext>,
    pub disposition: AttemptDisposition,
    pub outcome: RuntimeTerminalOutcome,
    pub primary_failure: Option<AttemptFailureSummary>,
    pub retry_advice: AttemptRetryAdvice,
    pub device_ordinal: Option<usize>,
    pub admission: Option<AttemptAdmissionFact>,
    pub admission_wait: Option<AttemptAdmissionWaitFact>,
    pub composed_schedule: Option<AttemptComposedScheduleProjection>,
    pub timestamps: AttemptTimestamps,
    pub cleanup_disposition: AttemptCleanupDisposition,
    pub cleanup_failure_count: u64,
    pub cleanup_failure_count_complete: bool,
    pub first_cleanup_failure: Option<CleanupFailureSummary>,
    pub capacity_disposition: AttemptCapacityDisposition,
    pub final_domain_usage: Option<AttemptFinalDomainUsage>,
    pub health_transition: Option<AttemptHealthTransition>,
    pub native_fault_evidence: Option<AttemptNativeFaultEvidence>,
    pub resource_evidence: Option<AttemptResourceEvidence>,
    pub observation_lost_records: u64,
    pub observation_retained_records: u64,
    pub observation_retained_bytes: u64,
    pub observation_truncated: bool,
    pub observation_degraded: bool,
    pub unix_epoch_ns: Option<u64>,
}

Immutable terminal attempt record published after cleanup and sealing.

The report preserves the execution outcome separately from cleanup, health and capacity disposition. Retry advice is derived from those committed facts; delivery failures do not change the report. Duration helpers return None for missing boundaries and saturate reversed intervals at zero.

Fields

schema: String

Schema identity accepted by report decoding.

attempt_id: AttemptId

Runtime identity shared by the attempt and its observations.

query_id: Option<AttemptQueryId>

Validated caller query correlation, or None when none was supplied.

trace_context: Option<AttemptTraceContext>

Validated opaque caller trace correlation, or None when none was supplied.

disposition: AttemptDisposition

Planning coverage of this attempt, independent of execution success.

outcome: RuntimeTerminalOutcome

First terminal execution outcome committed by the attempt owner.

primary_failure: Option<AttemptFailureSummary>

Bounded primary terminal-failure facts retained by the attempt lifecycle. Successful attempts have no primary summary.

retry_advice: AttemptRetryAdvice

The sole evidence-gated recovery recommendation derived as this report is sealed. It is required and never inferred from an error status at read time.

device_ordinal: Option<usize>

Selected CUDA device, or None when no device-backed admission occurred.

admission: Option<AttemptAdmissionFact>

Immutable grant facts, or None when the attempt was unadmitted.

admission_wait: Option<AttemptAdmissionWaitFact>

Final queue facts for an admitted request. None means the execution path did not retain a queue-backed admission handle; it is never reconstructed from snapshots after the report is sealed.

composed_schedule: Option<AttemptComposedScheduleProjection>

Immutable engine-owned composed schedule, when this admitted attempt used the composed graph execution contract. Ordinary attempts leave it absent; it is never reconstructed from runtime observations.

timestamps: AttemptTimestamps

Monotonic lifecycle boundary times for duration calculations.

cleanup_disposition: AttemptCleanupDisposition

Cleanup completion classification committed before publication.

cleanup_failure_count: u64

Exact when cleanup_failure_count_complete is true. The lifecycle never publishes a saturated arithmetic result as exact.

cleanup_failure_count_complete: bool

Required so decoding never fabricates an exact cleanup-failure count.

first_cleanup_failure: Option<CleanupFailureSummary>

First bounded cleanup failure. Further failures affect only the exact count; timeline retention is intentionally not authoritative.

capacity_disposition: AttemptCapacityDisposition

Disposition of the immutable grant after terminal cleanup.

final_domain_usage: Option<AttemptFinalDomainUsage>

Exact final domain usage after admitted cleanup, when sampling succeeded. None is an unavailable sample, never a zero sample.

health_transition: Option<AttemptHealthTransition>

Health result of the admitted release commit. None means no trustworthy admitted terminal release-commit fact was recorded (for example, an unadmitted attempt).

native_fault_evidence: Option<AttemptNativeFaultEvidence>

Bounded cuDF/cuGraph native status, isolation, and teardown facts from the same terminal release commit. None means the attempt did not own a device-backed domain. No native device status applies to such an attempt.

resource_evidence: Option<AttemptResourceEvidence>

Bounded ledger, immutable-cap, domain-usage, and allocator evidence from the sole terminal release commit. None means the attempt did not own a device-backed admission domain.

observation_lost_records: u64

Observation records lost because retention or recording failed.

observation_retained_records: u64

Records retained in the sealed bounded observation.

observation_retained_bytes: u64

Deterministic logical bytes retained in the sealed observation.

observation_truncated: bool

Whether retention limits prevented complete observation capture.

observation_degraded: bool

Whether recording degraded for reasons other than configured retention limits.

unix_epoch_ns: Option<u64>

Optional wall-clock nanosecond anchor associated with the observation clock.

Implementations

impl QueryAttemptReport

pub fn sealed_projection(&self) -> SealedAttemptProjection

Returns the compact, fixed-shape projection used by post-seal public consumers. The complete report remains the canonical retained record; this helper is only a read-only projection and never reconstructs facts from an error or delivery result.

impl QueryAttemptReport

pub fn total_ns(&self) -> Option<u64>

Returns submission-to-report-ready duration in nanoseconds. Returns None when report readiness was not recorded.

pub fn planning_ns(&self) -> Option<u64>

Returns planning-start-to-planning-finish duration in nanoseconds. Returns None when either boundary is absent.

pub fn queue_wait_ns(&self) -> Option<u64>

Returns admission-request-to-grant duration in nanoseconds. Returns None when either boundary is absent.

pub fn post_planning_pre_admission_ns(&self) -> Option<u64>

Returns planning-finish-to-admission-request duration in nanoseconds. Returns None when either boundary is absent.

pub fn admitted_to_physical_execution_ns(&self) -> Option<u64>

Returns grant-to-physical-execution-start duration in nanoseconds. Returns None when either boundary is absent.

pub fn physical_to_gpu_dispatch_ns(&self) -> Option<u64>

Returns physical-execution-start-to-GPU-dispatch duration in nanoseconds. Returns None when either boundary is absent.

pub fn admitted_to_gpu_dispatch_ns(&self) -> Option<u64>

Returns grant-to-first-GPU-dispatch duration in nanoseconds. Returns None when either boundary is absent.

pub fn submitted_to_gpu_dispatch_ns(&self) -> Option<u64>

Returns submission-to-first-GPU-dispatch duration in nanoseconds. Returns None when GPU dispatch was not recorded.

pub fn gpu_dispatch_to_terminal_ns(&self) -> Option<u64>

Returns first-GPU-dispatch-to-terminal-outcome duration in nanoseconds. Returns None when GPU dispatch was not recorded.

pub fn physical_execution_ns(&self) -> Option<u64>

Returns physical-execution-start-to-terminal-outcome duration in nanoseconds. Returns None when physical execution did not start.

pub fn cleanup_ns(&self) -> Option<u64>

Returns terminal-outcome-to-cleanup-complete duration in nanoseconds. Returns None when cleanup completion was not recorded.

Trait Implementations

impl Clone for QueryAttemptReport

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 QueryAttemptReport

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for QueryAttemptReport

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Eq for QueryAttemptReport

impl From<&QueryAttemptReport> for SealedAttemptProjection

fn from(report: &QueryAttemptReport) -> Self

Converts to this type from the input type.

impl PartialEq for QueryAttemptReport

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

Equality operator ==.

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

Inequality operator !=.

impl Serialize for QueryAttemptReport

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for QueryAttemptReport