Struct SegmentSizeStats 

#[non_exhaustive]
pub struct SegmentSizeStats {
    pub count: u64,
    pub min_bytes: u64,
    pub max_bytes: u64,
    pub mean_bytes: u64,
    pub p50_bytes: u64,
    pub p90_bytes: u64,
}

Size distribution of the on-disk segment files at a point in time.

Returned by SegmentBuffer::segment_size_stats. Unlike BufferStats (which derives BufferStats::segment_count and BufferStats::approx_disk_bytes from cheap atomic counters maintained on the flush/delete hot path), this struct is computed by a fresh directory scan: every field reflects the segment files as they actually are at call time. It is the tuning primitive for FlushPolicy::Batch: it answers “are my segments the size I expect, or is the batch size producing too many tiny files / too few huge ones?”

All byte values are the on-disk (compressed, post-envelope) file lengths, not item counts. Two segments holding the same number of items can differ in bytes because of compression and payload shape, so byte-size distribution is the honest signal for disk-footprint tuning.

Percentile definition

p50_bytes and p90_bytes use the nearest-rank method: the value returned is always an actual segment file size, never an interpolation between two. For n segments sorted ascending, the p-th percentile is the element at 1-based rank clamp(ceil(p / 100 · n), 1, n). Consequences:

• With one segment, min, p50, p90, and max are all equal.
• p50 is the lower median (the ceil(n / 2)-th smallest element).
• p90 is the size at or below which ~90% of segments fall.

When the buffer has no on-disk segments (nothing flushed yet, or everything acked), every field is 0.

This struct is #[non_exhaustive]: new fields (e.g. p99_bytes) may be added in any release without breaking semver.

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.

count: u64

Number of segment files on disk at scan time. Equals BufferStats::segment_count immediately after a sync_disk_bytes, but may differ from the live atomic counter between recalibrations.

min_bytes: u64

Smallest segment file size in bytes. 0 when there are no segments.

max_bytes: u64

Largest segment file size in bytes. 0 when there are no segments.

mean_bytes: u64

Arithmetic mean segment size (total_bytes / count), truncated to the integer. 0 when there are no segments.

p50_bytes: u64

Median (50th percentile) segment size, nearest-rank. 0 when there are no segments.

p90_bytes: u64

90th percentile segment size, nearest-rank. 0 when there are no segments.

Trait Implementations

impl Clone for SegmentSizeStats

fn clone(&self) -> SegmentSizeStats

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Copy for SegmentSizeStats

impl Debug for SegmentSizeStats

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

Formats the value using the given formatter.

impl Eq for SegmentSizeStats

impl PartialEq for SegmentSizeStats

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for SegmentSizeStats