Merge pull request #685 from getredash/fix/alert_sub_migration

Feature: add settings to query results cleanup

redash/models.py

@@ -369,10 +369,10 @@ def to_dict(self):
         }
 
     @classmethod
-    def unused(cls):
-        week_ago = datetime.datetime.now() - datetime.timedelta(days=7)
+    def unused(cls, days=7):
+        age_threshold = datetime.datetime.now() - datetime.timedelta(days=days)
 
-        unused_results = cls.select().where(Query.id == None, cls.retrieved_at < week_ago)\
+        unused_results = cls.select().where(Query.id == None, cls.retrieved_at < age_threshold)\
             .join(Query, join_type=peewee.JOIN_LEFT_OUTER)
 
         return unused_results

redash/settings.py

@@ -69,6 +69,8 @@ def all_settings():
 
 # The following enables periodic job (every 5 minutes) of removing unused query results.
 QUERY_RESULTS_CLEANUP_ENABLED = parse_boolean(os.environ.get("REDASH_QUERY_RESULTS_CLEANUP_ENABLED", "true"))
+QUERY_RESULTS_CLEANUP_COUNT = int(os.environ.get("REDASH_QUERY_RESULTS_CLEANUP_COUNT", "100"))
+QUERY_RESULTS_CLEANUP_MAX_AGE = int(os.environ.get("REDASH_QUERY_RESULTS_CLEANUP_MAX_AGE", "7"))
 
 AUTH_TYPE = os.environ.get("REDASH_AUTH_TYPE", "api_key")
 PASSWORD_LOGIN_ENABLED = parse_boolean(os.environ.get("REDASH_PASSWORD_LOGIN_ENABLED", "true"))

redash/tasks.py

@@ -221,7 +221,10 @@ def cleanup_query_results():
     Each time the job deletes only 100 query results so it won't choke the database in case of many such results.
     """
 
-    unused_query_results = models.QueryResult.unused().limit(100)
+    logging.info("Running query results clean up (removing maximum of %d unused results, that are %d days old or more)",
+                 settings.QUERY_RESULTS_CLEANUP_COUNT, settings.QUERY_RESULTS_CLEANUP_MAX_AGE)
+
+    unused_query_results = models.QueryResult.unused(settings.QUERY_RESULTS_CLEANUP_MAX_AGE).limit(settings.QUERY_RESULTS_CLEANUP_COUNT)
     total_unused_query_results = models.QueryResult.unused().count()
     deleted_count = models.QueryResult.delete().where(models.QueryResult.id << unused_query_results).execute()