Skip to content

AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.

fix(backend): Fix exception response on RPC error & Missing Graph `Running` Status (#9529)

Zamil Majdy538ad06547c0
diff --git a/autogpt_platform/backend/backend/data/execution.py b/autogpt_platform/backend/backend/data/execution.py
index bd6d79d47e..57dba68ab7 100644
--- a/autogpt_platform/backend/backend/data/execution.py
+++ b/autogpt_platform/backend/backend/data/execution.py
@@ -265,14 +265,18 @@ async def upsert_execution_output(
)
-async def update_graph_execution_start_time(graph_exec_id: str):
- await AgentGraphExecution.prisma().update(
+async def update_graph_execution_start_time(graph_exec_id: str) -> ExecutionResult:
+ res = await AgentGraphExecution.prisma().update(
where={"id": graph_exec_id},
data={
"executionStatus": ExecutionStatus.RUNNING,
"startedAt": datetime.now(tz=timezone.utc),
},
)
+ if not res:
+ raise ValueError(f"Execution {graph_exec_id} not found.")
+
+ return ExecutionResult.from_graph(res)
async def update_graph_execution_stats(
diff --git a/autogpt_platform/backend/backend/executor/database.py b/autogpt_platform/backend/backend/executor/database.py
index fad1a678de..5ead75c5ae 100644
--- a/autogpt_platform/backend/backend/executor/database.py
+++ b/autogpt_platform/backend/backend/executor/database.py
@@ -8,6 +8,7 @@ from backend.data.execution import (
get_incomplete_executions,
get_latest_execution,
update_execution_status,
+ update_graph_execution_start_time,
update_graph_execution_stats,
update_node_execution_stats,
upsert_execution_input,
@@ -52,6 +53,9 @@ class DatabaseManager(AppService):
get_incomplete_executions = exposed_run_and_wait(get_incomplete_executions)
get_latest_execution = exposed_run_and_wait(get_latest_execution)
update_execution_status = exposed_run_and_wait(update_execution_status)
+ update_graph_execution_start_time = exposed_run_and_wait(
+ update_graph_execution_start_time
+ )
update_graph_execution_stats = exposed_run_and_wait(update_graph_execution_stats)
update_node_execution_stats = exposed_run_and_wait(update_node_execution_stats)
upsert_execution_input = exposed_run_and_wait(upsert_execution_input)
diff --git a/autogpt_platform/backend/backend/executor/manager.py b/autogpt_platform/backend/backend/executor/manager.py
index 6ad8f05164..7dfcc86dca 100644
--- a/autogpt_platform/backend/backend/executor/manager.py
+++ b/autogpt_platform/backend/backend/executor/manager.py
@@ -644,6 +644,7 @@ class Executor:
node_eid="*",
block_name="-",
)
+ cls.db_client.update_graph_execution_start_time(graph_exec.graph_exec_id)
timing_info, (exec_stats, status, error) = cls._on_graph_execution(
graph_exec, cancel, log_metadata
)
diff --git a/autogpt_platform/backend/backend/util/service.py b/autogpt_platform/backend/backend/util/service.py
index 5f1c8b56f0..189b20b13c 100644
--- a/autogpt_platform/backend/backend/util/service.py
+++ b/autogpt_platform/backend/backend/util/service.py
@@ -36,7 +36,7 @@ from typing import (
import httpx
import Pyro5.api
import uvicorn
-from fastapi import FastAPI, HTTPException
+from fastapi import FastAPI, Request, responses
from pydantic import BaseModel, TypeAdapter, create_model
from Pyro5 import api as pyro
from Pyro5 import config as pyro_config
@@ -52,9 +52,11 @@ T = TypeVar("T")
C = TypeVar("C", bound=Callable)
config = Config()
-pyro_config.MAX_RETRIES = config.pyro_client_comm_retry # type: ignore
-pyro_config.COMMTIMEOUT = config.pyro_client_comm_timeout # type: ignore
api_host = config.pyro_host
+api_comm_retry = config.pyro_client_comm_retry
+api_comm_timeout = config.pyro_client_comm_timeout
+pyro_config.MAX_RETRIES = api_comm_retry # type: ignore
+pyro_config.COMMTIMEOUT = api_comm_timeout # type: ignore
P = ParamSpec("P")
@@ -241,6 +243,18 @@ class BaseAppService(AppProcess, ABC):
class FastApiAppService(BaseAppService, ABC):
fastapi_app: FastAPI
+ @staticmethod
+ def _handle_internal_http_error(status_code: int = 500, log_error: bool = True):
+ def handler(request: Request, exc: Exception):
+ if log_error:
+ logger.exception(f"{request.method} {request.url.path} failed: {exc}")
+ return responses.Response(
+ content=str(exc),
+ status_code=status_code,
+ )
+
+ return handler
+
def _create_fastapi_endpoint(self, func: Callable) -> Callable:
"""
Generates a FastAPI endpoint for the given function, handling default and optional parameters.
@@ -274,25 +288,15 @@ class FastApiAppService(BaseAppService, ABC):
if asyncio.iscoroutinefunction(f):
async def async_endpoint(body: RequestBodyModel): # type: ignore #RequestBodyModel being variable
- try:
- return await f(
- **{name: getattr(body, name) for name in body.model_fields}
- )
- except Exception as e:
- logger.exception(f"Error in {func.__name__}: {e}")
- raise HTTPException(status_code=500, detail=e)
+ return await f(
+ **{name: getattr(body, name) for name in body.model_fields}
+ )
return async_endpoint
else:
def sync_endpoint(body: RequestBodyModel): # type: ignore #RequestBodyModel being variable
- try:
- return f(
- **{name: getattr(body, name) for name in body.model_fields}
- )
- except Exception as e:
- logger.exception(f"Error in {func.__name__}: {e}")
- raise HTTPException(status_code=500, detail=e)
+ return f(**{name: getattr(body, name) for name in body.model_fields})
return sync_endpoint
@@ -322,6 +326,12 @@ class FastApiAppService(BaseAppService, ABC):
self.fastapi_app.add_api_route(
"/health_check", self.health_check, methods=["POST"]
)
+ self.fastapi_app.add_exception_handler(
+ ValueError, self._handle_internal_http_error(400)
+ )
+ self.fastapi_app.add_exception_handler(
+ Exception, self._handle_internal_http_error(500)
+ )
# Start the FastAPI server in a separate thread.
api_thread = threading.Thread(target=self.__start_fastapi, daemon=True)
@@ -393,19 +403,21 @@ def fastapi_close_service_client(client: Any) -> None:
logger.warning(f"Client {client} is not closable")
-@conn_retry("FastAPI client", "Creating service client")
+@conn_retry("FastAPI client", "Creating service client", max_retry=api_comm_retry)
def fastapi_get_service_client(service_type: Type[AS]) -> AS:
class DynamicClient:
def __init__(self):
host = service_type.get_host()
port = service_type.get_port()
self.base_url = f"http://{host}:{port}".rstrip("/")
- self.client = httpx.Client()
+ self.client = httpx.Client(
+ base_url=self.base_url,
+ timeout=api_comm_timeout,
+ )
def _call_method(self, method_name: str, **kwargs) -> Any:
try:
- url = f"{self.base_url}/{method_name}"
- response = self.client.post(url, json=to_dict(kwargs))
+ response = self.client.post(method_name, json=to_dict(kwargs))
response.raise_for_status()
return response.json()
except httpx.HTTPStatusError as e:
diff --git a/autogpt_platform/backend/backend/util/settings.py b/autogpt_platform/backend/backend/util/settings.py
index 1593b6a5a8..613093acd5 100644
--- a/autogpt_platform/backend/backend/util/settings.py
+++ b/autogpt_platform/backend/backend/util/settings.py
@@ -66,7 +66,7 @@ class Config(UpdateTrackingModel["Config"], BaseSettings):
description="Maximum number of workers to use for node execution within a single graph.",
)
use_http_based_rpc: bool = Field(
- default=False,
+ default=True,
description="Whether to use HTTP-based RPC for communication between services.",
)
pyro_host: str = Field(