reduced the places where we probe hourly tables

internal/tasks/inventoryHelpers.go

@@ -103,19 +103,9 @@ LIMIT ?
 			continue
 		}
 		probed := false
-		hasRows := count.Valid && count.Int64 > 0
+		// If count is known and >0, trust it; if NULL, accept optimistically to avoid heavy probes.
+		hasRows := !count.Valid || count.Int64 > 0
 		start := time.Now()
-		if !count.Valid {
-			probed = true
-			if ok, err := db.TableHasRows(ctx, dbConn, name); err == nil {
-				hasRows = ok
-			} else {
-				hasRows = false
-				if logger != nil {
-					logger.Debug("snapshot table probe failed", "table", name, "error", err)
-				}
-			}
-		}
 		if vcenter != "" && hasRows {
 			probed = true
 			vrows, qerr := querySnapshotRows(ctx, dbConn, name, []string{"VmId"}, `"Vcenter" = ? LIMIT 1`, vcenter)

internal/tasks/inventoryLifecycle.go

@@ -137,16 +137,10 @@ ORDER BY snapshot_time ASC
 		if err := db.ValidateTableName(t.Table); err != nil {
 			continue
 		}
-		if t.Count.Valid {
-			if t.Count.Int64 <= 0 {
+		// Trust snapshot_count if present; otherwise optimistically include to avoid long probes.
+		if t.Count.Valid && t.Count.Int64 <= 0 {
 			continue
 		}
-		} else {
-			hasRows, err := db.TableHasRows(ctx, dbConn, t.Table)
-			if err != nil || !hasRows {
-				continue
-			}
-		}
 		tables = append(tables, t)
 	}
 	return tables, nil