Struct GatewayConfig 

pub struct GatewayConfig {
    pub enabled: bool,
    pub top_n: usize,
    pub cache_ttl_secs: u64,
    pub call_timeout_secs: u64,
    pub servers: Vec<GatewayServer>,
    pub compress_output: bool,
    pub handle_spill: bool,
    pub index_output: bool,
    pub output_budget_tokens: usize,
}

[gateway] configuration block.

Fields

enabled: bool

Master switch. false → fully no-op (default).

top_n: usize

How many tools ctx_tools find returns per query.

cache_ttl_secs: u64

Aggregated-catalog cache lifetime (seconds).

call_timeout_secs: u64

Per-operation timeout for downstream connect/list/call (seconds).

servers: Vec<GatewayServer>

Downstream MCP servers to aggregate.

compress_output: bool

L1 (#1093): run downstream tool output through lean-ctx’s format-aware compressor before it reaches the model. false → output passes through unchanged (legacy). The transform is a deterministic function of (content, budget) so it never defeats provider prompt-caching (#498).

handle_spill: bool

L2 (#1094): when output exceeds output_budget_tokens, spill the verbatim blob to the content-addressed archive and hand the model a ctx_expand handle + summary instead of the full payload.

index_output: bool

L3 (#1095): side-channel — consolidate downstream output into the BM25 index, property graph, and knowledge store (so ctx_search / ctx_semantic_search find it later), without altering the returned text.

output_budget_tokens: usize

Token budget driving the L1 compression target and the L2 spill threshold. Inert while every post-processing flag is off.

Implementations

impl GatewayConfig

pub fn enabled_effective(&self) -> bool

Effective enabled flag, honoring the LEAN_CTX_GATEWAY env override (0|false|off disables, anything else enables).

pub fn active_servers(&self) -> impl Iterator<Item = &GatewayServer>

Enabled servers in declaration order.

pub fn effective_top_n(&self) -> usize

Clamp top_n into a sane range (1..=50).

pub fn postprocess_active(&self) -> bool

Whether any output post-processing is active (L1 compress / L2 spill / L3 index). When false, super::postprocess is a pure pass-through and the proxy hot-path pays nothing.

pub fn effective_output_budget(&self) -> usize

Effective output token budget, clamped away from the degenerate 0 (which would make L1 target nothing and L2 spill everything). Floors at 256 tokens so a misconfigured 0 still yields sane behaviour.

Trait Implementations

impl Clone for GatewayConfig

fn clone(&self) -> GatewayConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for GatewayConfig

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

Formats the value using the given formatter.

impl Default for GatewayConfig

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for GatewayConfig

where GatewayConfig: Default,

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Eq for GatewayConfig

impl PartialEq for GatewayConfig

fn eq(&self, other: &GatewayConfig) -> 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 GatewayConfig

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 GatewayConfig