R-772 (found live on 9202 with v0.286.0): a stopped probe container is seen on the next tick, not when the last healthy record's 5-minute interval runs out

Finding the container costs no network call, so it now runs before the interval check. Red-proof RP-D1b.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0159rPz1ZhFKsS53msqPYxtS

controller/internal/stacks/healthprobe.go

@@ -38,6 +38,16 @@ func (m *Manager) RunHealthProbes() error {
 			continue
 		}
 
+		// R-772 (measured live on 9202, v0.286.0): a container that stops is seen at once, not when the interval of the
+		// last HEALTHY result runs out (up to 5 minutes later) — finding the container costs no network call, so it is
+		// done before the interval check, and a stack with nothing to probe is recorded "not checked" on this tick.
+		containerName, candidates := findProbeContainerMeta(name, &stack.Meta, stack.Containers)
+		if containerName == "" {
+			skippedNoContainer++
+			noProbe = append(noProbe, noProbeStack{name: name, candidates: candidates})
+			continue
+		}
+
 		// Check if interval has elapsed since last probe.
 		// Fast 10s probes until healthy, then normal interval (default 5m).
 		// When HealthProbe is nil (just started/restarted), probe immediately.
@@ -58,16 +68,8 @@ func (m *Manager) RunHealthProbes() error {
 			}
 		}
 
-		// Find the main container to probe. R-630: a stack whose check resolves to no container
+		// R-630: a stack whose check resolves to no container used to be skipped SILENTLY — paperless-ngx's probe had
-		// used to be skipped SILENTLY — paperless-ngx's probe had therefore never run on any box,
+		// therefore never run on any box. It gets a RESULT that says why (above, before the interval check).
-		// and nothing on any screen could say so. It now gets a RESULT that says why, so the app
-		// page can tell "checked and healthy" from "never checked".
-		containerName, candidates := findProbeContainerMeta(name, &stack.Meta, stack.Containers)
-		if containerName == "" {
-			skippedNoContainer++
-			noProbe = append(noProbe, noProbeStack{name: name, candidates: candidates})
-			continue
-		}
 
 		targets = append(targets, probeTarget{
 			stackName:     name,

controller/internal/stacks/r772_not_checked_test.go

@@ -78,3 +78,16 @@ func TestProbeSaysUnhealthy_NotCheckedLeavesTheStateAlone(t *testing.T) {
 		t.Fatal("nil / healthy must not override")
 	}
 }
+
+// Measured live on 9202 (v0.286.0): with a HEALTHY record 2 minutes old, stopping the probe's container left the record
+// "healthy" until its 5-minute interval ran out. The missing container is now seen on the very next tick.
+func TestRunHealthProbes_AStoppedContainerIsSeenAtOnce(t *testing.T) {
+	m := r772Manager(time.Now())
+	st := r772Stack()
+	st.HealthProbe = &HealthProbeResult{Healthy: true, LastCheck: time.Now().Add(-2 * time.Minute)}
+	m.stacks["karakeep"] = st
+	_ = m.RunHealthProbes()
+	if hp := m.stacks["karakeep"].HealthProbe; hp.Healthy || !hp.NotChecked {
+		t.Fatalf("a stopped probe container must be recorded not checked on the next tick, got %+v", hp)
+	}
+}