feat(wal): expose runtime admission in node diagnostics

crates/lance-context-wal/src/service/health.rs

@@ -31,6 +31,7 @@ struct View {
     node: u64,
     ready: bool,
     groups: BTreeMap<u64, Group>,
+    admission: runtime::Admission,
     disk: Option<crate::admission::Snapshot>,
     maintenance: maintenance::Progress,
     snapshots_sent: u64,
@@ -63,6 +64,7 @@ impl State {
             node: self.node,
             ready,
             groups,
+            admission: self.handle.admission(),
             disk: self.handle.disk_budget(),
             maintenance: self
                 .maintenance

docs/design/wal-node-service.md

@@ -84,6 +84,11 @@ must eventually stop; checkpoints cannot discard that data to improve a score.
 ## Observation and qualification
 
 The diagnostic HTTP listener serves `/health/live`, `/health/ready` and `/status`.
+The status and readiness responses include `admission` observations for data,
+control, incoming peers and outgoing peers. Each lane reports request and byte
+capacity, current usage, and separate rejection counters. These independently
+sampled local values distinguish request pressure from response-byte pressure;
+they exclude the transport's separate budgets and do not change readiness.
 It allows at most 16 connections, bounds header buffers, disables keepalive and
 uses a two-second connection deadline, independently of WAL RPC capacity.
 Readiness reports active local workers and fresh local observations; it is not