feat(audit-log): surface partition-purge failure as health event + counter

src/ZB.MOM.WW.ScadaBridge.AuditLog/Central/IAuditCentralHealthSnapshot.cs

@@ -61,6 +61,16 @@ public interface IAuditCentralHealthSnapshot
     /// </summary>
     int AuditInboundCeilingHits { get; }
 
+    /// <summary>
+    /// Count of AuditLog partition-purge failures since process start. Incremented by
+    /// <see cref="AuditLogPurgeActor"/> every time a purge step throws — a partition
+    /// switch-out, a per-channel override DELETE, or the pre-purge boundary enumeration.
+    /// A sustained non-zero count means the retention job is silently not draining the
+    /// table; each failure also publishes an <see cref="AuditLogPurgeFailedEvent"/> on the
+    /// EventStream (arch-review 04, S2).
+    /// </summary>
+    int PurgeFailures { get; }
+
     /// <summary>
     /// Per-site latched stalled state: <c>true</c> when the
     /// <see cref="SiteAuditReconciliationActor"/> has observed two