feat(health): report permanently rejected site audit rows on the site health report (ScadaBridge#69)

SiteAuditBacklogSnapshot gains an optional trailing RejectedCount (gauge of
Rejected forward-state rows, read from the site store), carried on the wire as
SiteAuditBacklogSnapshotDto.rejected_count = 4 (additive; old sites read as 0,
old central ignores it). Health page shows an 'Audit rows rejected' badge.

Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV

tests/ZB.MOM.WW.ScadaBridge.AuditLog.Tests/Site/SqliteAuditWriterForwardStateTests.cs

@@ -133,6 +133,29 @@ public class SqliteAuditWriterForwardStateTests : IDisposable
         Assert.Equal(1, writer.RejectedPushRowCount);
     }
 
+    /// <summary>
+    /// The backlog snapshot reports permanently refused rows via <c>RejectedCount</c> (Gitea #69):
+    /// a Rejected row is still owed (counted in <c>PendingCount</c>) but is also counted separately,
+    /// and rows still Pending do not inflate it.
+    /// </summary>
+    /// <returns>A task that represents the asynchronous operation.</returns>
+    [Fact]
+    public async Task BacklogStats_ReportsRejectedCount_SeparatelyFromPending()
+    {
+        await using var writer = CreateWriter();
+        var t = DateTime.UtcNow.AddMinutes(-5);
+        var poison = await WriteAsync(writer, t);
+        await WriteAsync(writer, t.AddSeconds(1));
+
+        Assert.Equal(0, (await writer.GetBacklogStatsAsync()).RejectedCount);
+
+        await writer.RecordRejectedPushAttemptsAsync(new[] { poison }, maxAttempts: 1);
+
+        var stats = await writer.GetBacklogStatsAsync();
+        Assert.Equal(1, stats.RejectedCount);
+        Assert.Equal(2, stats.PendingCount); // Rejected is still owed to central via the pull
+    }
+
     /// <summary>
     /// The served high-water mark survives a writer restart: rows served before the restart are
     /// retired by the first cursor that covers them after it, instead of being stranded owed