feat(server): event-store health probe

Third HealthProbe (after disk + llama): times a cheap lastGlobalSequence()
read as an event-store responsiveness heartbeat — DEGRADED on throw or when
latency exceeds eventStoreLatencyWarnMs (default 500). Read-only by design
(no probe writes into the source-of-truth log; Invariant #1). Plugs into the
existing HealthMonitor; no new events. Observability spec §4.

Also wraps a long line in LlamaServerHealthProbeTest to clear a detekt
MaxLineLength finding introduced in aaf896d.

core/config/src/main/kotlin/com/correx/core/config/CorrexConfig.kt

@@ -29,6 +29,10 @@ data class CorrexConfig(
  * timeout within this window is reported as DEGRADED (liveness). [llamaTpsWarnBelow] is the
  * tokens/sec floor below which the probe reports degradation (KV-cache pressure, thermal throttle);
  * 0.0 disables the degradation dimension (no baseline to compare against on a fresh install).
+ *
+ * [eventStoreLatencyWarnMs] is the read-latency threshold above which the event-store probe reports
+ * DEGRADED. A single cheap read (lastGlobalSequence) is timed each tick; a slow or throwing read
+ * signals contention or unavailability. 500 ms suits typical SQLite on a homelab SSD.
  */
 @Serializable
 data class HealthConfig(
@@ -38,6 +42,7 @@ data class HealthConfig(
     val diskWarnGrowthBytesPerMin: Long = 50_000_000,
     val llamaLivenessTimeoutMs: Long = 5_000,
     val llamaTpsWarnBelow: Double = 0.0,
+    val eventStoreLatencyWarnMs: Long = 500,
 )
 
 /**