dbos 0.19.0a4__tar.gz → 0.20.0a2__tar.gz

{dbos-0.19.0a4 → dbos-0.20.0a2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: dbos
-Version: 0.19.0a4
+Version: 0.20.0a2
 Summary: Ultra-lightweight durable execution in Python
 Author-Email: "DBOS, Inc." <contact@dbos.dev>
 License: MIT
@@ -28,14 +28,14 @@ Description-Content-Type: text/markdown
 
 <div align="center">
 
-# DBOS Transact: Ultra-Lightweight Durable Execution
+# DBOS Transact: A Lightweight Durable Execution Library Built on Postgres
 
 #### [Documentation](https://docs.dbos.dev/) &nbsp;&nbsp;•&nbsp;&nbsp;  [Examples](https://docs.dbos.dev/examples) &nbsp;&nbsp;•&nbsp;&nbsp; [Github](https://github.com/dbos-inc) &nbsp;&nbsp;•&nbsp;&nbsp; [Discord](https://discord.com/invite/jsmC6pXGgX)
 </div>
 
 ---
 
-DBOS Transact is a Python library providing **ultra-lightweight durable execution**.
+DBOS Transact is a Python library for **ultra-lightweight durable execution**.
 For example:
 
 ```python
@@ -55,18 +55,23 @@ def workflow()
 
 Durable execution means your program is **resilient to any failure**.
 If it is ever interrupted or crashes, all your workflows will automatically resume from the last completed step.
-If you want to see durable execution in action, check out [this demo app](https://demo-widget-store.cloud.dbos.dev/) (source code [here](https://github.com/dbos-inc/dbos-demo-apps/tree/main/python/widget-store)).
-No matter how many times you try to crash it, it always resumes from exactly where it left off!
+Durable execution helps solve many common problems:
 
-Under the hood, DBOS Transact works by storing your program's execution state (which workflows are currently executing and which steps they've completed) in a Postgres database.
-So all you need to use it is a Postgres database to connect to&mdash;there's no need for a "workflow server."
-This approach is also incredibly fast, for example [25x faster than AWS Step Functions](https://www.dbos.dev/blog/dbos-vs-aws-step-functions-benchmark).
+- Orchestrating long-running or business-critical workflows so they seamlessly recover from any failure.
+- Running reliable background jobs with no timeouts.
+- Processing incoming events (e.g. from Kafka) exactly once.
+- Running a fault-tolerant distributed task queue.
+- Running a reliable cron scheduler.
+- Operating an AI agent, or anything that connects to an unreliable or non-deterministic API.
 
-Some more cool features include:
+What’s unique about DBOS's implementation of durable execution is that it’s implemented in a **lightweight library** that’s **totally backed by Postgres**.
+To use DBOS, just `pip install` it and annotate your program with DBOS decorators.
+Under the hood, those decorators store your program's execution state (which workflows are currently executing and which steps they've completed) in a Postgres database.
+If your program crashes or is interrupted, they automatically recover its workflows from their stored state.
+So all you need to use DBOS is Postgres&mdash;there are no other dependencies you have to manage, no separate workflow server.
 
-- Scheduled jobs&mdash;run your workflows exactly-once per time interval.
-- Exactly-once event processing&mdash;use workflows to process incoming events (for example, from a Kafka topic) exactly-once.
-- Observability&mdash;all workflows automatically emit [OpenTelemetry](https://opentelemetry.io/) traces.
+One big advantage of this approach is that you can add DBOS to **any** Python application&mdash;**it’s just a library**.
+You can use DBOS to add reliable background jobs or cron scheduling or queues to your app with no external dependencies except Postgres.
 
 ## Getting Started
 
@@ -77,7 +82,7 @@ pip install dbos
 dbos init --config
 ```
 
-Then, try it out with this simple program (requires Postgres):
+Then, try it out with this simple program:
 
 ```python
 from fastapi import FastAPI
@@ -107,14 +112,14 @@ def fastapi_endpoint():
     dbos_workflow()
 ```
 
-Save the program into `main.py`, edit `dbos-config.yaml` to configure your Postgres connection settings, and start it with `fastapi run`.
+Save the program into `main.py` and start it with `fastapi run`.
 Visit `localhost:8000` in your browser to start the workflow.
 When prompted, press `Control + \` to force quit your application.
 It should crash midway through the workflow, having completed step one but not step two.
 Then, restart your app with `fastapi run`.
 It should resume the workflow from where it left off, completing step two without re-executing step one.
 
-To learn how to build more complex workflows, see our [programming guide](https://docs.dbos.dev/python/programming-guide) or [examples](https://docs.dbos.dev/examples).
+To learn how to build more complex workflows, see the [programming guide](https://docs.dbos.dev/python/programming-guide) or [examples](https://docs.dbos.dev/examples).
 
 ## Documentation
 
@@ -125,7 +130,7 @@ To learn how to build more complex workflows, see our [programming guide](https:
 
 - [**AI-Powered Slackbot**](https://docs.dbos.dev/python/examples/rag-slackbot) &mdash; A Slackbot that answers questions about previous Slack conversations, using DBOS to durably orchestrate its RAG pipeline.
 - [**Widget Store**](https://docs.dbos.dev/python/examples/widget-store) &mdash; An online storefront that uses DBOS durable workflows to be resilient to any failure.
-- [**Earthquake Tracker**](https://docs.dbos.dev/python/examples/earthquake-tracker) &mdash; A real-time earthquake dashboard that uses DBOS to stream data from the USGS into Postgres, then visualizes it with Streamlit.
+- [**Scheduled Reminders**](https://docs.dbos.dev/python/examples/scheduled-reminders) &mdash; In just three lines of code, schedule an email to send days, weeks, or months in the future.
 
 More examples [here](https://docs.dbos.dev/examples)!
 

{dbos-0.19.0a4 → dbos-0.20.0a2}/README.md

@@ -1,14 +1,14 @@
 
 <div align="center">
 
-# DBOS Transact: Ultra-Lightweight Durable Execution
+# DBOS Transact: A Lightweight Durable Execution Library Built on Postgres
 
 #### [Documentation](https://docs.dbos.dev/) &nbsp;&nbsp;•&nbsp;&nbsp;  [Examples](https://docs.dbos.dev/examples) &nbsp;&nbsp;•&nbsp;&nbsp; [Github](https://github.com/dbos-inc) &nbsp;&nbsp;•&nbsp;&nbsp; [Discord](https://discord.com/invite/jsmC6pXGgX)
 </div>
 
 ---
 
-DBOS Transact is a Python library providing **ultra-lightweight durable execution**.
+DBOS Transact is a Python library for **ultra-lightweight durable execution**.
 For example:
 
 ```python
@@ -28,18 +28,23 @@ def workflow()
 
 Durable execution means your program is **resilient to any failure**.
 If it is ever interrupted or crashes, all your workflows will automatically resume from the last completed step.
-If you want to see durable execution in action, check out [this demo app](https://demo-widget-store.cloud.dbos.dev/) (source code [here](https://github.com/dbos-inc/dbos-demo-apps/tree/main/python/widget-store)).
-No matter how many times you try to crash it, it always resumes from exactly where it left off!
+Durable execution helps solve many common problems:
 
-Under the hood, DBOS Transact works by storing your program's execution state (which workflows are currently executing and which steps they've completed) in a Postgres database.
-So all you need to use it is a Postgres database to connect to&mdash;there's no need for a "workflow server."
-This approach is also incredibly fast, for example [25x faster than AWS Step Functions](https://www.dbos.dev/blog/dbos-vs-aws-step-functions-benchmark).
+- Orchestrating long-running or business-critical workflows so they seamlessly recover from any failure.
+- Running reliable background jobs with no timeouts.
+- Processing incoming events (e.g. from Kafka) exactly once.
+- Running a fault-tolerant distributed task queue.
+- Running a reliable cron scheduler.
+- Operating an AI agent, or anything that connects to an unreliable or non-deterministic API.
 
-Some more cool features include:
+What’s unique about DBOS's implementation of durable execution is that it’s implemented in a **lightweight library** that’s **totally backed by Postgres**.
+To use DBOS, just `pip install` it and annotate your program with DBOS decorators.
+Under the hood, those decorators store your program's execution state (which workflows are currently executing and which steps they've completed) in a Postgres database.
+If your program crashes or is interrupted, they automatically recover its workflows from their stored state.
+So all you need to use DBOS is Postgres&mdash;there are no other dependencies you have to manage, no separate workflow server.
 
-- Scheduled jobs&mdash;run your workflows exactly-once per time interval.
-- Exactly-once event processing&mdash;use workflows to process incoming events (for example, from a Kafka topic) exactly-once.
-- Observability&mdash;all workflows automatically emit [OpenTelemetry](https://opentelemetry.io/) traces.
+One big advantage of this approach is that you can add DBOS to **any** Python application&mdash;**it’s just a library**.
+You can use DBOS to add reliable background jobs or cron scheduling or queues to your app with no external dependencies except Postgres.
 
 ## Getting Started
 
@@ -50,7 +55,7 @@ pip install dbos
 dbos init --config
 ```
 
-Then, try it out with this simple program (requires Postgres):
+Then, try it out with this simple program:
 
 ```python
 from fastapi import FastAPI
@@ -80,14 +85,14 @@ def fastapi_endpoint():
     dbos_workflow()
 ```
 
-Save the program into `main.py`, edit `dbos-config.yaml` to configure your Postgres connection settings, and start it with `fastapi run`.
+Save the program into `main.py` and start it with `fastapi run`.
 Visit `localhost:8000` in your browser to start the workflow.
 When prompted, press `Control + \` to force quit your application.
 It should crash midway through the workflow, having completed step one but not step two.
 Then, restart your app with `fastapi run`.
 It should resume the workflow from where it left off, completing step two without re-executing step one.
 
-To learn how to build more complex workflows, see our [programming guide](https://docs.dbos.dev/python/programming-guide) or [examples](https://docs.dbos.dev/examples).
+To learn how to build more complex workflows, see the [programming guide](https://docs.dbos.dev/python/programming-guide) or [examples](https://docs.dbos.dev/examples).
 
 ## Documentation
 
@@ -98,7 +103,7 @@ To learn how to build more complex workflows, see our [programming guide](https:
 
 - [**AI-Powered Slackbot**](https://docs.dbos.dev/python/examples/rag-slackbot) &mdash; A Slackbot that answers questions about previous Slack conversations, using DBOS to durably orchestrate its RAG pipeline.
 - [**Widget Store**](https://docs.dbos.dev/python/examples/widget-store) &mdash; An online storefront that uses DBOS durable workflows to be resilient to any failure.
-- [**Earthquake Tracker**](https://docs.dbos.dev/python/examples/earthquake-tracker) &mdash; A real-time earthquake dashboard that uses DBOS to stream data from the USGS into Postgres, then visualizes it with Streamlit.
+- [**Scheduled Reminders**](https://docs.dbos.dev/python/examples/scheduled-reminders) &mdash; In just three lines of code, schedule an email to send days, weeks, or months in the future.
 
 More examples [here](https://docs.dbos.dev/examples)!
 

{dbos-0.19.0a4 → dbos-0.20.0a2}/dbos/_admin_server.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 import json
+import re
 import threading
 from functools import partial
 from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
@@ -15,6 +16,9 @@ if TYPE_CHECKING:
 _health_check_path = "/dbos-healthz"
 _workflow_recovery_path = "/dbos-workflow-recovery"
 _deactivate_path = "/deactivate"
+# /workflows/:workflow_id/cancel
+# /workflows/:workflow_id/resume
+# /workflows/:workflow_id/restart
 
 
 class AdminServer:
@@ -79,12 +83,51 @@ class AdminRequestHandler(BaseHTTPRequestHandler):
             self._end_headers()
             self.wfile.write(json.dumps(workflow_ids).encode("utf-8"))
         else:
-            self.send_response(404)
-            self._end_headers()
+
+            restart_match = re.match(
+                r"^/workflows/(?P<workflow_id>[^/]+)/restart$", self.path
+            )
+            resume_match = re.match(
+                r"^/workflows/(?P<workflow_id>[^/]+)/resume$", self.path
+            )
+            cancel_match = re.match(
+                r"^/workflows/(?P<workflow_id>[^/]+)/cancel$", self.path
+            )
+
+            if restart_match:
+                workflow_id = restart_match.group("workflow_id")
+                self._handle_restart(workflow_id)
+            elif resume_match:
+                workflow_id = resume_match.group("workflow_id")
+                self._handle_resume(workflow_id)
+            elif cancel_match:
+                workflow_id = cancel_match.group("workflow_id")
+                self._handle_cancel(workflow_id)
+            else:
+                self.send_response(404)
+                self._end_headers()
 
     def log_message(self, format: str, *args: Any) -> None:
         return  # Disable admin server request logging
 
+    def _handle_restart(self, workflow_id: str) -> None:
+        self.dbos.restart_workflow(workflow_id)
+        print("Restarting workflow", workflow_id)
+        self.send_response(204)
+        self._end_headers()
+
+    def _handle_resume(self, workflow_id: str) -> None:
+        print("Resuming workflow", workflow_id)
+        self.dbos.resume_workflow(workflow_id)
+        self.send_response(204)
+        self._end_headers()
+
+    def _handle_cancel(self, workflow_id: str) -> None:
+        print("Cancelling workflow", workflow_id)
+        self.dbos.cancel_workflow(workflow_id)
+        self.send_response(204)
+        self._end_headers()
+
 
 # Be consistent with DBOS-TS response.
 class PerfUtilization(TypedDict):

{dbos-0.19.0a4 → dbos-0.20.0a2}/dbos/_context.py

@@ -57,6 +57,7 @@ class DBOSContext:
         self.request: Optional["Request"] = None
 
         self.id_assigned_for_next_workflow: str = ""
+        self.is_within_set_workflow_id_block: bool = False
 
         self.parent_workflow_id: str = ""
         self.parent_workflow_fid: int = -1
@@ -78,6 +79,7 @@ class DBOSContext:
         rv.logger = self.logger
         rv.id_assigned_for_next_workflow = self.id_assigned_for_next_workflow
         self.id_assigned_for_next_workflow = ""
+        rv.is_within_set_workflow_id_block = self.is_within_set_workflow_id_block
         rv.parent_workflow_id = self.workflow_id
         rv.parent_workflow_fid = self.function_id
         rv.in_recovery = self.in_recovery
@@ -95,6 +97,10 @@ class DBOSContext:
         if len(self.id_assigned_for_next_workflow) > 0:
             wfid = self.id_assigned_for_next_workflow
         else:
+            if self.is_within_set_workflow_id_block:
+                self.logger.warning(
+                    f"Multiple workflows started in the same SetWorkflowID block. Only the first workflow is assigned the specified workflow ID; subsequent workflows will use a generated workflow ID."
+                )
             wfid = str(uuid.uuid4())
         return wfid
 
@@ -286,7 +292,7 @@ class DBOSContextSwap:
 
 class SetWorkflowID:
     """
-    Set the workflow ID to be used for the enclosed workflow invocation.
+    Set the workflow ID to be used for the enclosed workflow invocation. Note: Only the first workflow will be started with the specified workflow ID within a `with SetWorkflowID` block.
 
     Typical Usage
         ```
@@ -311,7 +317,9 @@ class SetWorkflowID:
         if ctx is None:
             self.created_ctx = True
             _set_local_dbos_context(DBOSContext())
-        assert_current_dbos_context().id_assigned_for_next_workflow = self.wfid
+        ctx = assert_current_dbos_context()
+        ctx.id_assigned_for_next_workflow = self.wfid
+        ctx.is_within_set_workflow_id_block = True
         return self
 
     def __exit__(
@@ -321,6 +329,7 @@ class SetWorkflowID:
         traceback: Optional[TracebackType],
     ) -> Literal[False]:
         # Code to clean up the basic context if we created it
+        assert_current_dbos_context().is_within_set_workflow_id_block = False
         if self.created_ctx:
             _clear_local_dbos_context()
         return False  # Did not handle

{dbos-0.19.0a4 → dbos-0.20.0a2}/dbos/_core.py

@@ -188,6 +188,7 @@ def _init_workflow(
         wf_status = dbos._sys_db.update_workflow_status(
             status, False, ctx.in_recovery, max_recovery_attempts=max_recovery_attempts
         )
+        # TODO: Modify the inputs if they were changed by `update_workflow_inputs`
         dbos._sys_db.update_workflow_inputs(wfid, _serialization.serialize_args(inputs))
     else:
         # Buffer the inputs for single-transaction workflows, but don't buffer the status
@@ -265,7 +266,9 @@ def _execute_workflow_wthread(
                 raise
 
 
-def execute_workflow_by_id(dbos: "DBOS", workflow_id: str) -> "WorkflowHandle[Any]":
+def execute_workflow_by_id(
+    dbos: "DBOS", workflow_id: str, startNew: bool = False
+) -> "WorkflowHandle[Any]":
     status = dbos._sys_db.get_workflow_status(workflow_id)
     if not status:
         raise DBOSRecoveryError(workflow_id, "Workflow status not found")
@@ -292,7 +295,8 @@ def execute_workflow_by_id(dbos: "DBOS", workflow_id: str) -> "WorkflowHandle[An
                     workflow_id,
                     f"Cannot execute workflow because instance '{iname}' is not registered",
                 )
-            with SetWorkflowID(workflow_id):
+
+            if startNew:
                 return start_workflow(
                     dbos,
                     wf_func,
@@ -302,6 +306,17 @@ def execute_workflow_by_id(dbos: "DBOS", workflow_id: str) -> "WorkflowHandle[An
                     *inputs["args"],
                     **inputs["kwargs"],
                 )
+            else:
+                with SetWorkflowID(workflow_id):
+                    return start_workflow(
+                        dbos,
+                        wf_func,
+                        status["queue_name"],
+                        True,
+                        dbos._registry.instance_info_map[iname],
+                        *inputs["args"],
+                        **inputs["kwargs"],
+                    )
         elif status["class_name"] is not None:
             class_name = status["class_name"]
             if class_name not in dbos._registry.class_info_map:
@@ -309,7 +324,8 @@ def execute_workflow_by_id(dbos: "DBOS", workflow_id: str) -> "WorkflowHandle[An
                     workflow_id,
                     f"Cannot execute workflow because class '{class_name}' is not registered",
                 )
-            with SetWorkflowID(workflow_id):
+
+            if startNew:
                 return start_workflow(
                     dbos,
                     wf_func,
@@ -319,8 +335,19 @@ def execute_workflow_by_id(dbos: "DBOS", workflow_id: str) -> "WorkflowHandle[An
                     *inputs["args"],
                     **inputs["kwargs"],
                 )
+            else:
+                with SetWorkflowID(workflow_id):
+                    return start_workflow(
+                        dbos,
+                        wf_func,
+                        status["queue_name"],
+                        True,
+                        dbos._registry.class_info_map[class_name],
+                        *inputs["args"],
+                        **inputs["kwargs"],
+                    )
         else:
-            with SetWorkflowID(workflow_id):
+            if startNew:
                 return start_workflow(
                     dbos,
                     wf_func,
@@ -329,6 +356,16 @@ def execute_workflow_by_id(dbos: "DBOS", workflow_id: str) -> "WorkflowHandle[An
                     *inputs["args"],
                     **inputs["kwargs"],
                 )
+            else:
+                with SetWorkflowID(workflow_id):
+                    return start_workflow(
+                        dbos,
+                        wf_func,
+                        status["queue_name"],
+                        True,
+                        *inputs["args"],
+                        **inputs["kwargs"],
+                    )
 
 
 @overload
@@ -422,6 +459,9 @@ def start_workflow(
         or wf_status == WorkflowStatusString.ERROR.value
         or wf_status == WorkflowStatusString.SUCCESS.value
     ):
+        dbos.logger.debug(
+            f"Workflow {new_wf_id} already completed with status {wf_status}. Directly returning a workflow handle."
+        )
         return WorkflowHandlePolling(new_wf_id, dbos)
 
     if fself is not None:
@@ -494,7 +534,7 @@ def workflow_wrapper(
                 temp_wf_type=get_temp_workflow_type(func),
                 max_recovery_attempts=max_recovery_attempts,
             )
-
+            # TODO: maybe modify the parameters if they've been changed by `_init_workflow`
             dbos.logger.debug(
                 f"Running workflow, id: {ctx.workflow_id}, name: {get_dbos_func_name(func)}"
             )

{dbos-0.19.0a4 → dbos-0.20.0a2}/dbos/_dbos.py

@@ -56,6 +56,7 @@ from ._registrations import (
 )
 from ._roles import default_required_roles, required_roles
 from ._scheduler import ScheduledWorkflow, scheduled
+from ._sys_db import WorkflowStatusString
 from ._tracer import dbos_tracer
 
 if TYPE_CHECKING:
@@ -231,6 +232,7 @@ class DBOS:
                         f"DBOS configured multiple times with conflicting information"
                     )
                 config = _dbos_global_registry.config
+
             _dbos_global_instance = super().__new__(cls)
             _dbos_global_instance.__init__(fastapi=fastapi, config=config, flask=flask)  # type: ignore
         else:
@@ -767,6 +769,11 @@ class DBOS:
         """Execute a workflow by ID (for recovery)."""
         return execute_workflow_by_id(_get_dbos_instance(), workflow_id)
 
+    @classmethod
+    def restart_workflow(cls, workflow_id: str) -> None:
+        """Execute a workflow by ID (for recovery)."""
+        execute_workflow_by_id(_get_dbos_instance(), workflow_id, True)
+
     @classmethod
     def recover_pending_workflows(
         cls, executor_ids: List[str] = ["local"]
@@ -774,6 +781,18 @@ class DBOS:
         """Find all PENDING workflows and execute them."""
         return recover_pending_workflows(_get_dbos_instance(), executor_ids)
 
+    @classmethod
+    def cancel_workflow(cls, workflow_id: str) -> None:
+        """Cancel a workflow by ID."""
+        _get_dbos_instance()._sys_db.set_workflow_status(
+            workflow_id, WorkflowStatusString.CANCELLED, False
+        )
+
+    @classmethod
+    def resume_workflow(cls, workflow_id: str) -> None:
+        """Resume a workflow by ID."""
+        execute_workflow_by_id(_get_dbos_instance(), workflow_id, False)
+
     @classproperty
     def logger(cls) -> Logger:
         """Return the DBOS `Logger` for the current context."""

{dbos-0.19.0a4 → dbos-0.20.0a2}/dbos/_error.py

@@ -35,6 +35,7 @@ class DBOSErrorCode(Enum):
     DeadLetterQueueError = 6
     MaxStepRetriesExceeded = 7
     NotAuthorized = 8
+    ConflictingWorkflowError = 9
 
 
 class DBOSWorkflowConflictIDError(DBOSException):
@@ -47,6 +48,16 @@ class DBOSWorkflowConflictIDError(DBOSException):
         )
 
 
+class DBOSConflictingWorkflowError(DBOSException):
+    """Exception raised different workflows started with the same workflow ID."""
+
+    def __init__(self, workflow_id: str, message: Optional[str] = None):
+        super().__init__(
+            f"Conflicting workflow invocation with the same ID ({workflow_id}): {message}",
+            dbos_error_code=DBOSErrorCode.ConflictingWorkflowError.value,
+        )
+
+
 class DBOSRecoveryError(DBOSException):
     """Exception raised when a workflow recovery fails."""
 

{dbos-0.19.0a4 → dbos-0.20.0a2}/dbos/_fastapi.py

@@ -94,7 +94,11 @@ def setup_fastapi_middleware(app: FastAPI, dbos: DBOS) -> None:
         with EnterDBOSHandler(attributes):
             ctx = assert_current_dbos_context()
             ctx.request = _make_request(request)
-            workflow_id = request.headers.get("dbos-idempotency-key", "")
-            with SetWorkflowID(workflow_id):
+            workflow_id = request.headers.get("dbos-idempotency-key")
+            if workflow_id is not None:
+                # Set the workflow ID for the handler
+                with SetWorkflowID(workflow_id):
+                    response = await call_next(request)
+            else:
                 response = await call_next(request)
         return response

{dbos-0.19.0a4 → dbos-0.20.0a2}/dbos/_flask.py

@@ -34,8 +34,12 @@ class FlaskMiddleware:
         with EnterDBOSHandler(attributes):
             ctx = assert_current_dbos_context()
             ctx.request = _make_request(request)
-            workflow_id = request.headers.get("dbos-idempotency-key", "")
-            with SetWorkflowID(workflow_id):
+            workflow_id = request.headers.get("dbos-idempotency-key")
+            if workflow_id is not None:
+                # Set the workflow ID for the handler
+                with SetWorkflowID(workflow_id):
+                    response = self.app(environ, start_response)
+            else:
                 response = self.app(environ, start_response)
         return response
 

{dbos-0.19.0a4 → dbos-0.20.0a2}/dbos/_kafka.py

@@ -1,3 +1,4 @@
+import re
 import threading
 from typing import TYPE_CHECKING, Any, Callable, NoReturn
 
@@ -19,6 +20,14 @@ _kafka_queue: Queue
 _in_order_kafka_queues: dict[str, Queue] = {}
 
 
+def safe_group_name(method_name: str, topics: list[str]) -> str:
+    safe_group_id = "-".join(
+        re.sub(r"[^a-zA-Z0-9\-]", "", str(r)) for r in [method_name, *topics]
+    )
+
+    return f"dbos-kafka-group-{safe_group_id}"[:255]
+
+
 def _kafka_consumer_loop(
     func: _KafkaConsumerWorkflow,
     config: dict[str, Any],
@@ -34,6 +43,12 @@ def _kafka_consumer_loop(
     if "auto.offset.reset" not in config:
         config["auto.offset.reset"] = "earliest"
 
+    if config.get("group.id") is None:
+        config["group.id"] = safe_group_name(func.__qualname__, topics)
+        dbos_logger.warning(
+            f"Consumer group ID not found. Using generated group.id {config['group.id']}"
+        )
+
     consumer = Consumer(config)
     try:
         consumer.subscribe(topics)
@@ -71,8 +86,9 @@ def _kafka_consumer_loop(
                     topic=cmsg.topic(),
                     value=cmsg.value(),
                 )
+                groupID = config.get("group.id")
                 with SetWorkflowID(
-                    f"kafka-unique-id-{msg.topic}-{msg.partition}-{msg.offset}"
+                    f"kafka-unique-id-{msg.topic}-{msg.partition}-{groupID}-{msg.offset}"
                 ):
                     if in_order:
                         assert msg.topic is not None

{dbos-0.19.0a4 → dbos-0.20.0a2}/dbos/_queue.py

@@ -36,6 +36,7 @@ class Queue:
         name: str,
         concurrency: Optional[int] = None,
         limiter: Optional[QueueRateLimit] = None,
+        *,  # Disable positional arguments from here on
         worker_concurrency: Optional[int] = None,
     ) -> None:
         if (