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

2026-07-09 06:49:28 -04:00
parent b59aa2d717
commit 14c4df54f9
9 changed files with 202 additions and 5 deletions
@@ -543,4 +543,50 @@ public class AuditLogPurgeActorTests : TestKit, IClassFixture<MsSqlMigrationFixt
            duration: TimeSpan.FromSeconds(3),
            interval: TimeSpan.FromMilliseconds(50));
    }
+
+    // ---------------------------------------------------------------------
+    // 11. SwitchThrows_PublishesPurgeFailedEvent_AndIncrementsHealthCounter (Task 5)
+    // ---------------------------------------------------------------------
+
+    [Fact]
+    public void SwitchThrows_PublishesPurgeFailedEvent_AndIncrementsHealthCounter()
+    {
+        // A silently failing retention job is a health-surface condition, not just a log line:
+        // the per-boundary catch must publish AuditLogPurgeFailedEvent AND bump the
+        // AuditCentralHealthSnapshot.PurgeFailures counter (arch-review 04, S2).
+        var boundary = new DateTime(2025, 9, 1, 0, 0, 0, DateTimeKind.Utc);
+        var repo = new RecordingRepo
+        {
+            Boundaries = new List<DateTime> { boundary },
+            ThrowOnBoundary = boundary,
+            BoundaryException = new InvalidOperationException("simulated switch timeout"),
+        };
+
+        var snapshot = new AuditCentralHealthSnapshot();
+        var services = new ServiceCollection();
+        services.AddScoped<IAuditLogRepository>(_ => repo);
+        services.AddSingleton(snapshot);
+        services.AddSingleton<IAuditPurgeFailureCounter>(snapshot);
+        var sp = services.BuildServiceProvider();
+
+        var probe = CreateTestProbe();
+        Sys.EventStream.Subscribe(probe.Ref, typeof(AuditLogPurgeFailedEvent));
+
+        Sys.ActorOf(Props.Create(() => new AuditLogPurgeActor(
+            sp,
+            Options.Create(FastTickOptions()),
+            Options.Create(new AuditLogOptions()),
+            NullLogger<AuditLogPurgeActor>.Instance)));
+
+        var evt = probe.ExpectMsg<AuditLogPurgeFailedEvent>(TimeSpan.FromSeconds(5));
+        Assert.Equal(boundary, evt.MonthBoundary);
+        Assert.Contains("simulated switch timeout", evt.Error);
+        Assert.True(evt.ElapsedMilliseconds >= 0);
+
+        AwaitAssert(
+            () => Assert.True(snapshot.PurgeFailures >= 1,
+                $"expected PurgeFailures >= 1, got {snapshot.PurgeFailures}"),
+            duration: TimeSpan.FromSeconds(3),
+            interval: TimeSpan.FromMilliseconds(50));
+    }
 }