Enum Recommendation 

pub enum Recommendation {
    VacuumNeeded {
        dead_tup_ratio: f64,
    },
    PartitionRecommended {
        reason: String,
    },
    PartitionCountIncrease {
        current: i32,
        suggested: i32,
    },
    Healthy,
}

Recommendation produced by recommend for a single table.

Precedence

When multiple rules would fire, recommend returns the highest-priority match per this strict ordering (highest first):

1. Recommendation::VacuumNeeded — bloat dominates everything; autovacuum lag is the most operationally urgent signal because dead tuples block index health and inflate disk usage.
2. Recommendation::PartitionRecommended — an unpartitioned table has crossed the row-count threshold; partitioning is structural work that should land before the table grows further.
3. Recommendation::PartitionCountIncrease — partitions exist but average row count per partition exceeds the threshold; expanding the partition count is incremental tuning.
4. Recommendation::Healthy — no rule fires.

JSON shape

The #[serde(tag = "kind", rename_all = "snake_case")] attribute produces internally-tagged JSON like {"kind":"vacuum_needed","dead_tup_ratio":0.42}. T10.2’s --format json path serialises a sorted vector of {table, recommendation} pairs; the snake_case tag keeps the machine-readable output ergonomic for shell scripts and dashboards.

Variants

VacuumNeeded

Dead-tuple ratio exceeded threshold_vacuum; operator should run VACUUM (or tune autovacuum).

Fields

dead_tup_ratio: f64

n_dead_tup / (n_live_tup + n_dead_tup). Always strictly greater than the threshold that triggered the variant.

PartitionRecommended

Unpartitioned table whose live row count exceeds threshold_partition_rows; operator should partition.

Fields

reason: String

Human-readable explanation including the row count and the threshold that fired the rule. Stable string format so --format json consumers can grep for substrings ("not partitioned", "threshold:").

PartitionCountIncrease

Partitioned table whose average rows-per-partition exceeds threshold_partition_rows; operator should expand the partition count.

Fields

current: i32

Current partition count.

suggested: i32

Suggested partition count — currently a simple doubling. Bounded by i32::saturating_mul so pathological inputs (e.g. 1.5B partitions) cap at i32::MAX rather than overflowing.

Healthy

No recommendation — table is within all thresholds.

Trait Implementations

impl Clone for Recommendation

fn clone(&self) -> Recommendation

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Recommendation

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

Formats the value using the given formatter.

impl PartialEq for Recommendation

fn eq(&self, other: &Recommendation) -> 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 Serialize for Recommendation

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 Recommendation