Struct Diagnostics 

pub struct Diagnostics {

    pub dependence_length: usize,
    pub effective_sample_size: usize,
    pub stationarity_ratio: f64,
    pub stationarity_ok: bool,
    pub projection_mismatch_q: f64,
    pub projection_mismatch_threshold: f64,
    pub projection_mismatch_ok: bool,
    pub top_quantiles: Option<Vec<TopQuantile>>,
    pub outlier_rate_baseline: f64,
    pub outlier_rate_sample: f64,
    pub outlier_asymmetry_ok: bool,
    pub discrete_mode: bool,
    pub timer_resolution_ns: f64,
    pub duplicate_fraction: f64,
    pub preflight_ok: bool,
    pub calibration_samples: usize,
    pub total_time_secs: f64,
    pub warnings: Vec<String>,
    pub quality_issues: Vec<QualityIssue>,
    pub preflight_warnings: Vec<PreflightWarningInfo>,
    pub seed: Option<u64>,
    pub attacker_model: Option<String>,
    pub threshold_ns: f64,
    pub timer_name: String,
    pub platform: String,
    pub timer_fallback_reason: Option<String>,
    pub gibbs_iters_total: usize,
    pub gibbs_burnin: usize,
    pub gibbs_retained: usize,
    pub lambda_mean: f64,
    pub lambda_sd: f64,
    pub lambda_cv: f64,
    pub lambda_ess: f64,
    pub lambda_mixing_ok: bool,
    pub kappa_mean: f64,
    pub kappa_sd: f64,
    pub kappa_cv: f64,
    pub kappa_ess: f64,
    pub kappa_mixing_ok: bool,
}

Diagnostic information for debugging and analysis.

See spec Section 4.1 (Result Types).

Fields

dependence_length: usize

Block size used for bootstrap (Politis-White automatic selection).

effective_sample_size: usize

Effective sample size accounting for autocorrelation (ESS approx n / dependence_length).

stationarity_ratio: f64

Non-stationarity: ratio of inference to calibration variance. Values 0.5-2.0 are normal; >5.0 indicates non-stationarity.

stationarity_ok: bool

True if stationarity ratio is within acceptable bounds (0.5-2.0).

projection_mismatch_q: f64

Projection mismatch Q statistic.

Measures how well the 2D (shift, tail) model fits the 9D quantile differences. A high value indicates the effect is concentrated at specific quantiles rather than following the shift+tail pattern.

projection_mismatch_threshold: f64

Bootstrap-calibrated threshold for projection mismatch Q statistic. Q > threshold indicates the 2D projection may be unreliable.

projection_mismatch_ok: bool

True if projection fits well (Q <= threshold).

When false, the shift_ns and tail_ns estimates should be interpreted with caution; use top_quantiles for more detailed information.

top_quantiles: Option<Vec<TopQuantile>>

Top quantiles by exceedance probability (when projection mismatch detected).

When projection_mismatch_ok is false, this field contains detailed information about which quantiles drive the leak detection. This helps diagnose effects that don’t fit the shift+tail model (e.g., effects concentrated at a single quantile).

outlier_rate_baseline: f64

Outlier rate for baseline class (fraction trimmed).

outlier_rate_sample: f64

Outlier rate for sample class (fraction trimmed).

outlier_asymmetry_ok: bool

True if outlier rates are symmetric (both <1%, ratio <3x, diff <2%).

discrete_mode: bool

Whether discrete timer mode was used (low timer resolution).

timer_resolution_ns: f64

Timer resolution in nanoseconds.

duplicate_fraction: f64

Fraction of samples with duplicate timing values (0.0-1.0).

preflight_ok: bool

True if preflight checks passed (sanity, generator, system).

calibration_samples: usize

Number of samples used for calibration (covariance estimation).

total_time_secs: f64

Total time spent on the analysis in seconds.

warnings: Vec<String>

Human-readable warnings (empty if all checks pass).

quality_issues: Vec<QualityIssue>

Quality issues detected during measurement.

preflight_warnings: Vec<PreflightWarningInfo>

Preflight warnings from calibration phase.

These warnings are categorized by severity:

• Informational: Sampling efficiency issues (results still valid)
• ResultUndermining: Statistical assumption violations (results may be unreliable)

seed: Option<u64>

Measurement seed used for reproducibility.

attacker_model: Option<String>

Attacker model name (e.g., “AdjacentNetwork”, “SharedHardware”).

threshold_ns: f64

Effect threshold (theta) in nanoseconds.

timer_name: String

Timer implementation name (e.g., “rdtsc”, “cntvct_el0”, “kperf”).

platform: String

Platform description (e.g., “macos-aarch64”).

timer_fallback_reason: Option<String>

Reason the timer fell back from high-precision PMU (if applicable).

Used to generate context-aware recommendations in output.

• “concurrent access”: kperf locked by another process
• “no sudo”: not running with elevated privileges
• “unavailable”: PMU init failed despite privileges
• None: using high-precision timer or x86_64 (rdtsc is already ~0.3ns)

gibbs_iters_total: usize

v5.4: Total number of Gibbs iterations.

gibbs_burnin: usize

v5.4: Number of burn-in iterations.

gibbs_retained: usize

v5.4: Number of retained samples.

lambda_mean: f64

v5.4: Posterior mean of latent scale λ.

lambda_sd: f64

v5.4: Posterior standard deviation of λ.

lambda_cv: f64

v5.4: Coefficient of variation of λ (λ_sd / λ_mean).

lambda_ess: f64

v5.4: Effective sample size of λ chain.

lambda_mixing_ok: bool

v5.4: Whether λ chain mixed well (CV ≥ 0.1 AND ESS ≥ 20).

kappa_mean: f64

v5.6: Posterior mean of likelihood precision κ.

kappa_sd: f64

v5.6: Posterior standard deviation of κ.

kappa_cv: f64

v5.6: Coefficient of variation of κ (kappa_sd / kappa_mean).

kappa_ess: f64

v5.6: Effective sample size of κ chain.

kappa_mixing_ok: bool

v5.6: Whether κ chain mixed well (CV ≥ 0.1 AND ESS ≥ 20).

Implementations

impl Diagnostics

pub fn all_ok() -> Diagnostics

Create diagnostics indicating all checks passed.

Uses placeholder values for numeric fields; prefer constructing explicitly with actual measured values.

pub fn all_checks_passed(&self) -> bool

Check if all diagnostics are OK.

Trait Implementations

impl Clone for Diagnostics

fn clone(&self) -> Diagnostics

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Diagnostics

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

Formats the value using the given formatter.

impl Default for Diagnostics

fn default() -> Diagnostics

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for Diagnostics

fn deserialize<__D>( __deserializer: __D, ) -> Result<Diagnostics, <__D as Deserializer<'de>>::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for Diagnostics

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

where __S: Serializer,

Serialize this value into the given Serde serializer.