fix: добавил max_failed_age — игнорировать старые failed runs (>24ч)

- failed runs старше 24 часов пропускаются как историческая ошибка
- предотвращает массовый перезапуск DAG-ов от 2024 года
- параметр max_failed_age настраивается в config.yaml

airflow_monitor/analyzer.py

@@ -27,11 +27,13 @@ class DagIssue:
 class DagAnalyzer:
     """Analyzes DAG runs and produces a list of issues."""
 
-    def __init__(self, long_running_threshold: int):
+    def __init__(self, long_running_threshold: int, max_failed_age: int = 86400):
         self._threshold = long_running_threshold
+        self._max_failed_age = max_failed_age  # ignore failed runs older than this (seconds)
         logger.debug(
-            "DagAnalyzer initialized: long_running_threshold=%ds (%.1f min)",
+            "DagAnalyzer initialized: long_running_threshold=%ds (%.1f min), max_failed_age=%ds (%.1f hours)",
             long_running_threshold, long_running_threshold / 60,
+            max_failed_age, max_failed_age / 3600,
         )
 
     def analyze_dag_runs(
@@ -81,6 +83,14 @@ class DagAnalyzer:
             )
 
             if state == "failed":
+                # Skip old failed runs — only alert on recent failures
+                if duration > self._max_failed_age:
+                    logger.debug(
+                        "  → SKIP: failed run too old (%.0fs > %ds max_failed_age)",
+                        duration, self._max_failed_age,
+                    )
+                    continue
+
                 error_info = self._extract_error(dag_id, run_id, task_instances_fn)
                 issue = DagIssue(
                     dag_id=dag_id,

airflow_monitor/config.py

@@ -25,6 +25,7 @@ class MonitorConfig:
 
     cycle_interval: int = 300  # seconds between full cycles
     long_running_threshold: int = 1800  # 30 minutes
+    max_failed_age: int = 86400  # ignore failed runs older than 24 hours
     retry_wait: int = 120  # wait after retry before re-check
     max_retries: int = 1
     state_file: str = "/var/lib/airflow-monitor/state.json"

airflow_monitor/monitor.py

@@ -32,7 +32,10 @@ class Monitor:
         self._shutdown = shutdown_event
         self._client = AirflowClient(config.airflow)
         self._state = StateManager(config.monitor.state_file, config.monitor.state_max_age)
-        self._analyzer = DagAnalyzer(config.monitor.long_running_threshold)
+        self._analyzer = DagAnalyzer(
+            config.monitor.long_running_threshold,
+            config.monitor.max_failed_age,
+        )
         self._exporter = DataExporter(config.zabbix)
         self._actions = ActionHandler(
             self._client, self._state, self._exporter, config.monitor,

config.yaml

@@ -36,6 +36,10 @@ monitor:
   # DAG run duration threshold to consider as "long running" (seconds)
   long_running_threshold: 1800  # 30 minutes
 
+  # Игнорировать failed runs старше этого порога (seconds)
+  # Старые ошибки — это история, не текущая проблема
+  max_failed_age: 86400  # 24 hours
+
   # Time to wait after retry before re-checking (seconds)
   retry_wait: 120  # 2 minutes