Platform for stateful agents: AI with advanced memory that can learn and self-improve over time.
feat: add gating around agent run cancellation (#3446)
cthomasc85dca4ec854
diff --git a/letta/server/rest_api/routers/v1/agents.py b/letta/server/rest_api/routers/v1/agents.pyindex 07fbcee57..55f5099ec 100644--- a/letta/server/rest_api/routers/v1/agents.py+++ b/letta/server/rest_api/routers/v1/agents.py@@ -705,28 +705,32 @@ async def send_message(model_compatible = agent.llm_config.model_endpoint_type in ["anthropic", "openai", "together", "google_ai", "google_vertex", "bedrock"]# Create a new run for execution tracking- job_status = JobStatus.created- run = await server.job_manager.create_job_async(- pydantic_job=Run(- user_id=actor.id,- status=job_status,- metadata={- "job_type": "send_message",- "agent_id": agent_id,- },- request_config=LettaRequestConfig(- use_assistant_message=request.use_assistant_message,- assistant_message_tool_name=request.assistant_message_tool_name,- assistant_message_tool_kwarg=request.assistant_message_tool_kwarg,- include_return_message_types=request.include_return_message_types,+ if settings.track_agent_run:+ job_status = JobStatus.created+ run = await server.job_manager.create_job_async(+ pydantic_job=Run(+ user_id=actor.id,+ status=job_status,+ metadata={+ "job_type": "send_message",+ "agent_id": agent_id,+ },+ request_config=LettaRequestConfig(+ use_assistant_message=request.use_assistant_message,+ assistant_message_tool_name=request.assistant_message_tool_name,+ assistant_message_tool_kwarg=request.assistant_message_tool_kwarg,+ include_return_message_types=request.include_return_message_types,+ ),),- ),- actor=actor,- )+ actor=actor,+ )+ else:+ run = None+job_update_metadata = None# TODO (cliandy): clean this upredis_client = await get_redis_client()- await redis_client.set(f"{REDIS_RUN_ID_PREFIX}:{agent_id}", run.id)+ await redis_client.set(f"{REDIS_RUN_ID_PREFIX}:{agent_id}", run.id if run else None)try:if agent_eligible and model_compatible:@@ -741,7 +745,7 @@ async def send_message(job_manager=server.job_manager,actor=actor,group=agent.multi_agent_group,- current_run_id=run.id,+ current_run_id=run.id if run else None,)else:agent_loop = LettaAgent(@@ -754,7 +758,7 @@ async def send_message(actor=actor,step_manager=server.step_manager,telemetry_manager=server.telemetry_manager if settings.llm_api_logging else NoopTelemetryManager(),- current_run_id=run.id,+ current_run_id=run.id if run else None,# summarizer settings to be added heresummarizer_mode=(SummarizationMode.STATIC_MESSAGE_BUFFER@@ -790,12 +794,13 @@ async def send_message(job_status = JobStatus.failedraisefinally:- await server.job_manager.safe_update_job_status_async(- job_id=run.id,- new_status=job_status,- actor=actor,- metadata=job_update_metadata,- )+ if settings.track_agent_run:+ await server.job_manager.safe_update_job_status_async(+ job_id=run.id,+ new_status=job_status,+ actor=actor,+ metadata=job_update_metadata,+ )# noinspection PyInconsistentReturns@@ -836,29 +841,32 @@ async def send_message_streaming(not_letta_endpoint = agent.llm_config.model_endpoint != LETTA_MODEL_ENDPOINT# Create a new job for execution tracking- job_status = JobStatus.created- run = await server.job_manager.create_job_async(- pydantic_job=Run(- user_id=actor.id,- status=job_status,- metadata={- "job_type": "send_message_streaming",- "agent_id": agent_id,- },- request_config=LettaRequestConfig(- use_assistant_message=request.use_assistant_message,- assistant_message_tool_name=request.assistant_message_tool_name,- assistant_message_tool_kwarg=request.assistant_message_tool_kwarg,- include_return_message_types=request.include_return_message_types,+ if settings.track_agent_run:+ job_status = JobStatus.created+ run = await server.job_manager.create_job_async(+ pydantic_job=Run(+ user_id=actor.id,+ status=job_status,+ metadata={+ "job_type": "send_message_streaming",+ "agent_id": agent_id,+ },+ request_config=LettaRequestConfig(+ use_assistant_message=request.use_assistant_message,+ assistant_message_tool_name=request.assistant_message_tool_name,+ assistant_message_tool_kwarg=request.assistant_message_tool_kwarg,+ include_return_message_types=request.include_return_message_types,+ ),),- ),- actor=actor,- )+ actor=actor,+ )+ else:+ run = Nonejob_update_metadata = None# TODO (cliandy): clean this upredis_client = await get_redis_client()- await redis_client.set(f"{REDIS_RUN_ID_PREFIX}:{agent_id}", run.id)+ await redis_client.set(f"{REDIS_RUN_ID_PREFIX}:{agent_id}", run.id if run else None)try:if agent_eligible and model_compatible:@@ -875,7 +883,7 @@ async def send_message_streaming(step_manager=server.step_manager,telemetry_manager=server.telemetry_manager if settings.llm_api_logging else NoopTelemetryManager(),group=agent.multi_agent_group,- current_run_id=run.id,+ current_run_id=run.id if run else None,)else:agent_loop = LettaAgent(@@ -888,7 +896,7 @@ async def send_message_streaming(actor=actor,step_manager=server.step_manager,telemetry_manager=server.telemetry_manager if settings.llm_api_logging else NoopTelemetryManager(),- current_run_id=run.id,+ current_run_id=run.id if run else None,# summarizer settings to be added heresummarizer_mode=(SummarizationMode.STATIC_MESSAGE_BUFFER@@ -941,12 +949,13 @@ async def send_message_streaming(job_status = JobStatus.failedraisefinally:- await server.job_manager.safe_update_job_status_async(- job_id=run.id,- new_status=job_status,- actor=actor,- metadata=job_update_metadata,- )+ if settings.track_agent_run:+ await server.job_manager.safe_update_job_status_async(+ job_id=run.id,+ new_status=job_status,+ actor=actor,+ metadata=job_update_metadata,+ )@router.post("/{agent_id}/messages/cancel", operation_id="cancel_agent_run")@@ -963,6 +972,8 @@ async def cancel_agent_run("""actor = await server.user_manager.get_actor_or_default_async(actor_id=actor_id)+ if not settings.track_agent_run:+ raise HTTPException(status_code=400, detail="Agent run tracking is disabled")if not run_ids:redis_client = await get_redis_client()run_id = await redis_client.get(f"{REDIS_RUN_ID_PREFIX}:{agent_id}")diff --git a/letta/server/rest_api/routers/v1/jobs.py b/letta/server/rest_api/routers/v1/jobs.pyindex 90c108d93..2a3ec4648 100644--- a/letta/server/rest_api/routers/v1/jobs.py+++ b/letta/server/rest_api/routers/v1/jobs.py@@ -7,6 +7,7 @@ from letta.schemas.enums import JobStatusfrom letta.schemas.job import Jobfrom letta.server.rest_api.utils import get_letta_serverfrom letta.server.server import SyncServer+from letta.settings import settingsrouter = APIRouter(prefix="/jobs", tags=["jobs"])@@ -93,6 +94,8 @@ async def cancel_job(agent execution to terminate as soon as possible."""actor = await server.user_manager.get_actor_or_default_async(actor_id=actor_id)+ if not settings.track_agent_run:+ raise HTTPException(status_code=400, detail="Agent run tracking is disabled")try:# First check if the job exists and is in a cancellable statediff --git a/letta/settings.py b/letta/settings.pyindex 39fb754a2..e438faa21 100644--- a/letta/settings.py+++ b/letta/settings.py@@ -235,6 +235,7 @@ class Settings(BaseSettings):track_last_agent_run: bool = Field(default=False, description="Update last agent run metrics")track_errored_messages: bool = Field(default=True, description="Enable tracking for errored messages")track_stop_reason: bool = Field(default=True, description="Enable tracking stop reason on steps.")+ track_agent_run: bool = Field(default=True, description="Enable tracking agent run with cancellation support")# FastAPI Application Settingsuvicorn_workers: int = 1