feat: add timezone support for CRON triggers (#438)

docs/docs/exosphere/triggers.md

@@ -61,13 +61,15 @@ Define triggers in your graph template:
     {
       "type": "CRON",
       "value": {
-        "expression": "0 9 * * 1-5"
+        "expression": "0 9 * * 1-5",
+        "timezone": "America/New_York"
       }
     },
     {
       "type": "CRON",
       "value": {
-        "expression": "0 0 * * 0"
+        "expression": "0 0 * * 0",
+        "timezone": "UTC"
       }
     }
   ],
@@ -77,6 +79,8 @@ Define triggers in your graph template:
 }
 ```
 
+**Note:** The `timezone` field is optional and defaults to `"UTC"` if not specified. Use IANA timezone names (e.g., `"America/New_York"`, `"Europe/London"`, `"Asia/Tokyo"`).
+
 ### Python SDK Example
 
 ```python
@@ -109,8 +113,8 @@ async def create_scheduled_graph():
 
     # Define triggers for automatic execution
     triggers = [
-        CronTrigger(expression="0 2 * * *"),    # Daily at 2:00 AM
-        CronTrigger(expression="0 */4 * * *")   # Every 4 hours
+        CronTrigger(expression="0 2 * * *", timezone="America/New_York"),  # Daily at 2:00 AM EST/EDT
+        CronTrigger(expression="0 */4 * * *", timezone="UTC")              # Every 4 hours UTC
     ]
 
     # Create the graph with triggers
@@ -158,7 +162,7 @@ asyncio.run(create_scheduled_graph())
 
 1. **Avoid Peak Times**: Schedule resource-intensive workflows during off-peak hours
 2. **Stagger Executions**: If you have multiple graphs, stagger their execution times
-3. **Consider Time Zones**: Cron expressions use server time (UTC by default)
+3. **Consider Time Zones**: Specify the `timezone` parameter to ensure your cron expressions run at the correct local time. If not specified, defaults to UTC.
 4. **Resource Planning**: Ensure your infrastructure can handle scheduled workloads
 
 ### Error Handling
@@ -191,11 +195,31 @@ result = await state_manager.upsert_graph(
 )
 ```
 
+## Timezone Support
+
+Triggers now support specifying a timezone for cron expressions, allowing you to schedule jobs in your local timezone:
+
+```python
+# Schedule a report to run at 9 AM New York time (handles DST automatically)
+CronTrigger(expression="0 9 * * 1-5", timezone="America/New_York")
+
+# Schedule a job at 5 PM London time
+CronTrigger(expression="0 17 * * *", timezone="Europe/London")
+
+# Schedule using UTC (default)
+CronTrigger(expression="0 12 * * *", timezone="UTC")
+```
+
+**Important Notes:**
+- Use IANA timezone names (e.g., `"America/New_York"`, `"Europe/London"`, `"Asia/Tokyo"`)
+- Timezones automatically handle Daylight Saving Time (DST) transitions
+- If no timezone is specified, defaults to `"UTC"`
+- All trigger times are internally stored in UTC for consistency
+
 ## Limitations
 
 - **CRON Only**: Currently only cron-based scheduling is supported
 - **No Manual Override**: Scheduled executions cannot be manually cancelled once triggered
-- **Time Zone**: All cron expressions are evaluated in server time (UTC)
 - **Minimum Interval**: Avoid scheduling more frequently than every minute
 
 ## Next Steps

python-sdk/exospherehost/_version.py

@@ -1 +1 @@
-version = "0.0.3b1"
+version = "0.0.3b2"

python-sdk/exospherehost/models.py

@@ -160,4 +160,5 @@ def validate_default_values(cls, v: dict[str, str]) -> dict[str, str]:
         return normalized_dict
 
 class CronTrigger(BaseModel):
-    expression: str = Field(..., description="Cron expression for scheduling automatic graph execution. Uses standard 5-field format: minute hour day-of-month month day-of-week. Example: '0 9 * * 1-5' for weekdays at 9 AM.")
+    expression: str = Field(..., description="Cron expression for scheduling automatic graph execution. Uses standard 5-field format: minute hour day-of-month month day-of-week. Example: '0 9 * * 1-5' for weekdays at 9 AM.")
+    timezone: str = Field(default="UTC", description="Timezone for the cron expression (e.g., 'America/New_York', 'Europe/London', 'UTC'). Defaults to 'UTC'.")

python-sdk/exospherehost/statemanager.py

@@ -173,7 +173,8 @@ async def upsert_graph(self, graph_name: str, graph_nodes: list[GraphNodeModel],
                 {
                     "type": "CRON",
                     "value": {
-                        "expression": trigger.expression
+                        "expression": trigger.expression,
+                        "timezone": trigger.timezone
                     }
                 }
                 for trigger in triggers

state-manager/app/controller/upsert_graph_template.py

@@ -66,7 +66,7 @@ async def upsert_graph_template(namespace_name: str, graph_name: str, body: Upse
                 DatabaseTriggers.graph_name == graph_name,
                 DatabaseTriggers.trigger_status == TriggerStatusEnum.PENDING,
                 DatabaseTriggers.type == TriggerTypeEnum.CRON,
-                In(DatabaseTriggers.expression, [trigger.value["expression"] for trigger in old_triggers if trigger.type == TriggerTypeEnum.CRON])
+                In(DatabaseTriggers.expression, [trigger.value["expression"] for trigger in old_triggers if trigger.value.type == TriggerTypeEnum.CRON])
             ).delete_many()
 
         background_tasks.add_task(verify_graph, graph_template)

state-manager/app/models/db/trigger.py

@@ -9,6 +9,7 @@
 class DatabaseTriggers(Document):
     type: TriggerTypeEnum = Field(..., description="Type of the trigger")
     expression: Optional[str] = Field(default=None, description="Expression of the trigger")
+    timezone: Optional[str] = Field(default="UTC", description="Timezone for the trigger")
     graph_name: str = Field(..., description="Name of the graph")
     namespace: str = Field(..., description="Namespace of the graph")
     trigger_time: datetime = Field(..., description="Trigger time of the trigger")

state-manager/app/models/trigger_models.py

@@ -1,7 +1,11 @@
-from pydantic import BaseModel, Field, field_validator, model_validator
+from pydantic import BaseModel, Field, field_validator
 from enum import Enum
 from croniter import croniter
-from typing import Self
+from typing import Union, Annotated, Literal
+from zoneinfo import available_timezones
+
+# Cache available timezones at module level to avoid repeated filesystem queries
+_AVAILABLE_TIMEZONES = available_timezones()
 
 class TriggerTypeEnum(str, Enum):
     CRON = "CRON"
@@ -14,7 +18,9 @@ class TriggerStatusEnum(str, Enum):
     TRIGGERING = "TRIGGERING"
 
 class CronTrigger(BaseModel):
+    type: Literal[TriggerTypeEnum.CRON] = Field(default=TriggerTypeEnum.CRON, description="Type of the trigger")
     expression: str = Field(..., description="Cron expression for the trigger")
+    timezone: str = Field(default="UTC", description="Timezone for the cron expression (e.g., 'America/New_York', 'Europe/London', 'UTC')")
 
     @field_validator("expression")
     @classmethod
@@ -23,14 +29,24 @@ def validate_expression(cls, v: str) -> str:
             raise ValueError("Invalid cron expression")
         return v
 
+    @field_validator("timezone")
+    @classmethod
+    def validate_timezone(cls, v: str) -> str:
+        if v not in _AVAILABLE_TIMEZONES:
+            raise ValueError(f"Invalid timezone: {v}. Must be a valid IANA timezone (e.g., 'America/New_York', 'Europe/London', 'UTC')")
+        return v
+
+# Union type for all trigger types - add new trigger types here
+TriggerValue = Annotated[Union[CronTrigger], Field(discriminator="type")]
+
 class Trigger(BaseModel):
-    type: TriggerTypeEnum = Field(..., description="Type of the trigger")
-    value: dict = Field(default_factory=dict, description="Value of the trigger")
-
-    @model_validator(mode="after")
-    def validate_trigger(self) -> Self:
-        if self.type == TriggerTypeEnum.CRON:
-            CronTrigger.model_validate(self.value)
-        else:
-            raise ValueError(f"Unsupported trigger type: {self.type}")
-        return self
+    """
+    Extensible trigger model using discriminated unions.
+    To add a new trigger type:
+    1. Add the enum value to TriggerTypeEnum
+    2. Create a new trigger class (e.g., WebhookTrigger) with type field
+    3. Add it to the TriggerValue Union
+
+    Note: Access trigger type via trigger.value.type
+    """
+    value: TriggerValue = Field(..., description="Value of the trigger")

state-manager/app/tasks/trigger_cron.py

@@ -8,9 +8,13 @@
 from pymongo import ReturnDocument
 from pymongo.errors import DuplicateKeyError
 from app.config.settings import get_settings
+from zoneinfo import ZoneInfo
 import croniter
 import asyncio
 
+# Cache UTC timezone at module level to avoid repeated instantiation
+UTC = ZoneInfo("UTC")
+
 logger = LogsManager().get_logger()
 
 async def get_due_triggers(cron_time: datetime) -> DatabaseTriggers | None:
@@ -42,15 +46,26 @@ async def mark_as_failed(trigger: DatabaseTriggers):
 
 async def create_next_triggers(trigger: DatabaseTriggers, cron_time: datetime):
     assert trigger.expression is not None
-    iter = croniter.croniter(trigger.expression, trigger.trigger_time)
+
+    # Use the trigger's timezone, defaulting to UTC if not specified
+    tz = ZoneInfo(trigger.timezone or "UTC")
+
+    # Convert trigger_time to the specified timezone for croniter
+    trigger_time_tz = trigger.trigger_time.replace(tzinfo=UTC).astimezone(tz)
+    iter = croniter.croniter(trigger.expression, trigger_time_tz)
 
     while True:
-        next_trigger_time = iter.get_next(datetime)
+        # Get next trigger time in the specified timezone
+        next_trigger_time_tz = iter.get_next(datetime)
+
+        # Convert back to UTC for storage
+        next_trigger_time = next_trigger_time_tz.astimezone(UTC).replace(tzinfo=None)
 
         try:
             await DatabaseTriggers(
                 type=TriggerTypeEnum.CRON,
                 expression=trigger.expression,
+                timezone=trigger.timezone,
                 graph_name=trigger.graph_name,
                 namespace=trigger.namespace,
                 trigger_time=next_trigger_time,

state-manager/app/tasks/verify_graph.py

@@ -3,6 +3,7 @@
 
 from datetime import datetime
 from json_schema_to_pydantic import create_model
+from zoneinfo import ZoneInfo
 
 from app.models.db.graph_template_model import GraphTemplate
 from app.models.graph_template_validation_status import GraphTemplateValidationStatus
@@ -11,6 +12,9 @@
 from app.models.trigger_models import TriggerStatusEnum, TriggerTypeEnum
 from app.models.db.trigger import DatabaseTriggers
 
+# Cache UTC timezone at module level to avoid repeated instantiation
+UTC = ZoneInfo("UTC")
+
 logger = LogsManager().get_logger()
 
 async def verify_node_exists(graph_template: GraphTemplate, registered_nodes: list[RegisteredNode]) -> list[str]:
@@ -101,20 +105,36 @@ async def verify_inputs(graph_template: GraphTemplate, registered_nodes: list[Re
     return errors
 
 async def create_crons(graph_template: GraphTemplate):
-    expressions_to_create = set([trigger.value["expression"] for trigger in graph_template.triggers if trigger.type == TriggerTypeEnum.CRON])
+    # Build a map of (expression, timezone) -> CronTrigger for deduplication
+    triggers_to_create = {}
+    for trigger in graph_template.triggers:
+        if trigger.value.type == TriggerTypeEnum.CRON:
+            # trigger.value is already a validated CronTrigger instance
+            cron_trigger = trigger.value
+            triggers_to_create[(cron_trigger.expression, cron_trigger.timezone)] = cron_trigger
+
+    current_time = datetime.now(UTC).replace(tzinfo=None)
 
-    current_time = datetime.now()
-
     new_db_triggers = []
-    for expression in expressions_to_create:
-        iter = croniter.croniter(expression, current_time)
+    for (expression, timezone), cron_trigger in triggers_to_create.items():
+        # Use the validated timezone (guaranteed to be valid IANA timezone, never None)
+        tz = ZoneInfo(timezone)
+
+        # Get current time in the specified timezone
+        current_time_tz = current_time.replace(tzinfo=UTC).astimezone(tz)
+        iter = croniter.croniter(expression, current_time_tz)
+
+        # Get next trigger time in the specified timezone
+        next_trigger_time_tz = iter.get_next(datetime)
+
+        # Convert back to UTC for storage (remove timezone info for storage)
+        next_trigger_time = next_trigger_time_tz.astimezone(UTC).replace(tzinfo=None)
 
-        next_trigger_time = iter.get_next(datetime)
-
         new_db_triggers.append(
             DatabaseTriggers(
                 type=TriggerTypeEnum.CRON,
-                expression=expression,
+                expression=cron_trigger.expression,
+                timezone=cron_trigger.timezone,
                 graph_name=graph_template.name,
                 namespace=graph_template.namespace,
                 trigger_status=TriggerStatusEnum.PENDING,