From 72659853092accfbdb098c40ec0b1cc1b9339fdc Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E2=80=9CNaeel=E2=80=9D?= <tazet@narod.ru>
Date: Mon, 18 May 2026 13:27:50 +0400
Subject: [PATCH] fix(reconciler): health-check Active NS every 60s to restore
 deleted SA/RoleBindings

RunReconciler now runs two tickers:
- 30s: retry Failed namespaces (existing behavior)
- 60s: DispatchResync on Active namespaces; since registerNamespace is
  idempotent this is a no-op when SA/RoleBindings are intact and
  silently restores them if deleted

Fixes P1-A from integration test 2026-05-18: SA deleted from Active NS
was not being restored because reconciler only processed Failed NS.
---
 pkg/utils/namespace_manager.go | 39 ++++++++++++++++++++++++++++------
 1 file changed, 32 insertions(+), 7 deletions(-)

diff --git a/pkg/utils/namespace_manager.go b/pkg/utils/namespace_manager.go
index 435bbbc3..af1a17e9 100644
--- a/pkg/utils/namespace_manager.go
+++ b/pkg/utils/namespace_manager.go
@@ -74,7 +74,9 @@ type NamespaceManager interface {
 	MarkPartActive(namespace string, part string) (NamespaceRecord, bool)
 	MarkPartFailed(namespace string, part string, err error) (NamespaceRecord, bool)
 	Remove(name string) bool
-	// RunReconciler periodically retries namespaces stuck in NamespacePhaseFailed.
+	// RunReconciler periodically retries namespaces stuck in NamespacePhaseFailed (every 30 s)
+	// and performs a health-check on Active namespaces (every 60 s) by calling DispatchResync,
+	// which re-ensures SA/RoleBindings and executor-type registration are intact.
 	// Must be started as a goroutine; exits when ctx is cancelled.
 	// logger is used to report retry attempts and outcomes; pass zap.NewNop() to silence.
 	RunReconciler(ctx context.Context, logger *zap.Logger)
@@ -612,22 +614,27 @@ func (m *inMemoryNamespaceManager) Remove(name string) bool {
 	return true
 }
 
-// RunReconciler periodically finds namespaces in NamespacePhaseFailed and retries them
-// via DispatchResync. This ensures transient k8s API errors (e.g. temporary 503 on
-// EnsureNamespaceSA or executor-type AddNamespace) do not permanently strand a namespace.
+// RunReconciler periodically:
+//   - retries namespaces stuck in NamespacePhaseFailed via DispatchResync (every 30 s)
+//   - health-checks Active namespaces by calling DispatchResync (every 60 s); since
+//     registerNamespace is idempotent, this is a no-op when SA/RoleBindings are intact
+//     and silently restores them if they were deleted.
+//
 // logger receives one log line per retry attempt and per outcome.
 // Exits when ctx is cancelled.
 func (m *inMemoryNamespaceManager) RunReconciler(ctx context.Context, logger *zap.Logger) {
 	if logger == nil {
 		logger = zap.NewNop()
 	}
-	ticker := time.NewTicker(30 * time.Second)
-	defer ticker.Stop()
+	failedTicker := time.NewTicker(30 * time.Second)
+	defer failedTicker.Stop()
+	activeTicker := time.NewTicker(60 * time.Second)
+	defer activeTicker.Stop()
 	for {
 		select {
 		case <-ctx.Done():
 			return
-		case <-ticker.C:
+		case <-failedTicker.C:
 			m.mu.RLock()
 			var failedNS []string
 			for ns, rec := range m.records {
@@ -655,6 +662,24 @@ func (m *inMemoryNamespaceManager) RunReconciler(ctx context.Context, logger *za
 					logger.Info("namespace reconciler: resync succeeded", zap.String("namespace", ns))
 				}
 			}
+		case <-activeTicker.C:
+			m.mu.RLock()
+			var activeNS []string
+			for ns, rec := range m.records {
+				if rec.Phase == NamespacePhaseActive {
+					activeNS = append(activeNS, ns)
+				}
+			}
+			m.mu.RUnlock()
+			for _, ns := range activeNS {
+				_, _, err := m.DispatchResync(ctx, ns)
+				if err != nil {
+					logger.Warn("namespace reconciler: active NS health-check failed, marking failed for retry",
+						zap.String("namespace", ns),
+						zap.Error(err),
+					)
+				}
+			}
 		}
 	}
 }