add config setting for execution schedule timeout

CHANGELOG.rst

@@ -17,6 +17,7 @@ Added
 * Add ``get_entrypoint()`` method to ``ActionResourceManager`` attribute of st2client.
   #4791
 * Add support for orquesta task retry. (new feature)
+* Add config option ``scheduler.execution_scheduling_timeout_threshold_min``. Fixes #4887
 
 Changed
 ~~~~~~~

conf/st2.conf.sample

@@ -279,6 +279,8 @@ pool_size = 10
 retry_wait_msec = 3000
 # How often (in seconds) to look for zombie execution requests before rescheduling them.
 gc_interval = 10
+# How often in minutes to check for unscheduled actions
+execution_scheduling_timeout_threshold_min = 1
 
 [schema]
 # Version of JSON schema to use.
@@ -379,4 +381,3 @@ retry_max_jitter_msec = 1000
 retry_wait_fixed_msec = 1000
 # Max seconds to allow workflow execution be idled before it is identified as orphaned and cancelled by the garbage collector. A value of zero means the feature is disabled. This is disabled by default.
 gc_max_idle_sec = 0
-

st2actions/st2actions/scheduler/config.py

@@ -50,6 +50,9 @@ def _register_service_opts():
             default='/etc/st2/logging.scheduler.conf',
             help='Location of the logging configuration file.'
         ),
+        cfg.FloatOpt(
+            'execution_scheduling_timeout_threshold_min', default=1,
+            help='How long GC to search back in minutes for orphaned scheduled actions'),
         cfg.IntOpt(
             'pool_size', default=10,
             help='The size of the pool used by the scheduler for scheduling executions.'),

st2actions/st2actions/scheduler/handler.py

@@ -50,7 +50,6 @@
 # (< 5 seconds). If an item is still being marked as processing it likely indicates that the
 # scheduler process which was processing that item crashed or similar so we need to mark it as
 # "handling=False" so some other scheduler process can pick it up.
-EXECUTION_SCHEDUELING_TIMEOUT_THRESHOLD_MS = (60 * 1000)
 
 # When a policy delayed execution is detected it will be try to be rescheduled by the scheduler
 # again in this amount of milliseconds.
@@ -62,6 +61,9 @@ def __init__(self):
         self.message_type = LiveActionDB
         self._shutdown = False
         self._pool = eventlet.GreenPool(size=cfg.CONF.scheduler.pool_size)
+        self._execution_scheduling_timeout_threshold_ms = \
+                cfg.CONF.scheduler.execution_scheduling_timeout_threshold_min \
+                * 60 * 1000
         self._coordinator = coordination_service.get_coordinator(start_heart=True)
         self._main_thread = None
         self._cleanup_thread = None
@@ -100,7 +102,7 @@ def _reset_handling_flag(self):
         query = {
             'scheduled_start_timestamp__lte': date.append_milliseconds_to_time(
                 date.get_datetime_utc_now(),
-                -EXECUTION_SCHEDUELING_TIMEOUT_THRESHOLD_MS
+                -self._execution_scheduling_timeout_threshold_ms
             ),
             'handling': True
         }

st2tests/st2tests/config.py

@@ -283,6 +283,9 @@ def _register_ssh_runner_opts():
 
 def _register_scheduler_opts():
     scheduler_opts = [
+        cfg.FloatOpt(
+            'execution_scheduling_timeout_threshold_min', default=1,
+            help='How long GC to search back in minutes for orphaned scheduled actions'),
         cfg.IntOpt(
             'pool_size', default=10,
             help='The size of the pool used by the scheduler for scheduling executions.'),