fix(engine): six correctness fixes for the state machine
CI / test (pull_request) Successful in 1m59s
CI / lint (pull_request) Successful in 1m17s
CI / vulncheck (pull_request) Successful in 1m1s

1. Group auto-pause trap: remove the one-way Paused=true mutation
   from checkGroup — monitorRoutine skipped paused groups, so they
   could never re-evaluate or auto-unpause.

2. Retry logic: apply MaxRetries to all →DOWN transitions, not just
   UP→DOWN. New monitors (PENDING) no longer alert on first transient
   failure when retries are configured.

3. Shutdown drain hole: track checker goroutines with checkerWG so
   Stop() waits for in-flight checks before draining the write queue.
   Final drainWrites() catches any writes enqueued after the writer's
   own drain.

4. Probe-ingest writer bypass: route SaveCheckFromNode through the
   engine's serialized dbWriter instead of writing directly to the
   store from the HTTP handler.

5. Dead-probe expiry: expire stale probe results (>3× site interval)
   before aggregation so a dead probe can't poison status forever.
   Also clean probeResults in RemoveSite.

6. Maintenance-cache N+1: replace per-check DB query with a
   fully-resolved in-memory cache refreshed every poll cycle. One
   GetActiveMaintenanceWindows() call instead of N IsMonitorInMaintenance.

ImportData now wipes check_history, state_changes, and alert_health
so re-inserted IDs don't inherit stale history from prior occupants.
This commit was merged in pull request #105.
This commit is contained in:
2026-06-11 13:40:31 -04:00
parent 61c28fac62
commit 5d5153351e
7 changed files with 335 additions and 39 deletions
+40
View File
@@ -235,6 +235,46 @@ func TestImportExport(t *testing.T) {
}
}
func TestImportData_WipesHistory(t *testing.T) {
s := newTestStore(t)
if err := s.AddSite(models.Site{Name: "OldSite", URL: "https://old.com", Type: "http", Interval: 30}); err != nil {
t.Fatalf("AddSite: %v", err)
}
if err := s.SaveCheck(1, 5000, true); err != nil {
t.Fatalf("SaveCheck: %v", err)
}
if err := s.SaveStateChange(1, "UP", "DOWN", "timeout"); err != nil {
t.Fatalf("SaveStateChange: %v", err)
}
if err := s.SaveAlertHealth(models.AlertHealthRecord{AlertID: 1, LastSendOK: true, SendCount: 1}); err != nil {
t.Fatalf("SaveAlertHealth: %v", err)
}
backup := models.Backup{
Sites: []models.Site{{ID: 1, Name: "NewSite", URL: "https://new.com", Type: "http", Interval: 60}},
}
if err := s.ImportData(backup); err != nil {
t.Fatalf("ImportData: %v", err)
}
history, err := s.LoadAllHistory(100)
if err != nil {
t.Fatalf("LoadAllHistory: %v", err)
}
if len(history) != 0 {
t.Errorf("expected empty check_history after import, got %d sites with history", len(history))
}
changes, err := s.GetStateChanges(1, 100)
if err != nil {
t.Fatalf("GetStateChanges: %v", err)
}
if len(changes) != 0 {
t.Errorf("expected empty state_changes after import, got %d", len(changes))
}
}
func TestCheckHistory(t *testing.T) {
s := newTestStore(t)