Enum ScaleLadder 

pub enum ScaleLadder {
    Ns,
    Us,
    Bytes,
    Ticks,
    Unitless,
    None,
}

Closed enumeration of auto-scale ladders driving format dispatch.

Picks the unit family up the type system rather than a free-form &'static str tag. Each AggRule variant maps to exactly one ladder via AggRule::ladder; each super::DerivedMetricDef entry carries a ladder via super::DerivedMetricDef::ladder; the cgroup- level render path passes a ladder directly. A registry typo or drift between accessor newtype and ladder choice fails to compile at the registry edit site rather than silently routing through an “unknown unit” pass-through arm at render time.

The six ladder variants and their step-up rules:

• Ns: ns → µs (×1e3) → ms (×1e6) → s (×1e9). Decimal prefixes — SI time, not binary. Used for AggRule::SumNs (cumulative ns counters), AggRule::MaxPeak (lifetime ns high-water marks), AggRule::MaxGaugeNs (instantaneous ns gauges), and the "ns" derived-metric ladder.
• Us: µs → ms (×1e3) → s (×1e6). Decimal SI prefixes. The cgroup cpu_usage_usec and throttled_usec fields are reported by the kernel in microseconds; this ladder scales them up the same way the Ns ladder scales nanoseconds.
• Bytes: B → KiB → MiB → GiB → TiB. IEC binary prefixes (×1024) for byte counts. Used for AggRule::SumBytes, AggRule::MaxPeakBytes, and any byte-typed derived metric.
• Ticks: ticks → Kticks (×1e3) → Mticks (×1e6). Decimal prefixes for clock-tick counts (utime_clock_ticks, stime_clock_ticks); the unit itself is opaque (the kernel’s USER_HZ rate is host-dependent), so an SI prefix is the most we can promise.
• Unitless: “” → K → M → G. Decimal prefixes for non-dimensional counters (wakeups, migrations, csw, syscall counts). Used for AggRule::SumCount and AggRule::MaxGaugeCount.
• None: no ladder — values render as the bare integer with no unit suffix and no scaling. Used for AggRule::Mode / AggRule::ModeChar / AggRule::ModeBool (categorical strings), AggRule::RangeI32 / AggRule::RangeU32 (bounded ordinals), and AggRule::Affinity (cpuset summaries) — the Aggregated std::fmt::Display impl handles render for these directly.

The threshold for stepping up is |value| >= next_scale. Sign is preserved through scaling (negative deltas pass through). Zero stays at base unit.

Variants

Ns

Us

Bytes

Ticks

Unitless

None

Implementations

impl ScaleLadder

pub fn base_unit(&self) -> &'static str

Base unit string for this ladder — what auto_scale returns for a value at the bottom of the ladder. Used by the format helpers to detect whether a value stepped up (auto_scale(v).1 != ladder.base_unit() ⇒ stepped up, render with the scaled unit; equal ⇒ no step-up, render the bare integer with the base unit suffix).

Trait Implementations

impl Clone for ScaleLadder

fn clone(&self) -> ScaleLadder

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ScaleLadder

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

Formats the value using the given formatter.

impl Hash for ScaleLadder

fn hash<__H: Hasher>(&self, state: &mut __H)

Feeds this value into the given Hasher.

fn hash_slice<H>(data: &[Self], state: &mut H)

where H: Hasher, Self: Sized,

Feeds a slice of this type into the given Hasher.

impl PartialEq for ScaleLadder

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Copy for ScaleLadder

impl Eq for ScaleLadder

impl StructuralPartialEq for ScaleLadder