feat(api): add scheduled cleanup task for specific workflow logs (#31843)

Co-authored-by: 章润喆 <zhangrunzhe@zhangrunzhedeMacBook-Air.local>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: hjlarry <hjlarry@163.com>
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: hj24 <mambahj24@gmail.com>

api/.env.example

@@ -553,6 +553,8 @@ WORKFLOW_LOG_CLEANUP_ENABLED=false
 WORKFLOW_LOG_RETENTION_DAYS=30
 # Batch size for workflow log cleanup operations (default: 100)
 WORKFLOW_LOG_CLEANUP_BATCH_SIZE=100
+# Comma-separated list of workflow IDs to clean logs for
+WORKFLOW_LOG_CLEANUP_SPECIFIC_WORKFLOW_IDS=
 
 # App configuration
 APP_MAX_EXECUTION_TIME=1200
@@ -715,6 +717,7 @@ ANNOTATION_IMPORT_MAX_CONCURRENT=5
 # Sandbox expired records clean configuration
 SANDBOX_EXPIRED_RECORDS_CLEAN_GRACEFUL_PERIOD=21
 SANDBOX_EXPIRED_RECORDS_CLEAN_BATCH_SIZE=1000
+SANDBOX_EXPIRED_RECORDS_CLEAN_BATCH_MAX_INTERVAL=200
 SANDBOX_EXPIRED_RECORDS_RETENTION_DAYS=30
 SANDBOX_EXPIRED_RECORDS_CLEAN_TASK_LOCK_TTL=90000
 

api/configs/feature/__init__.py

@@ -1314,6 +1314,9 @@ class WorkflowLogConfig(BaseSettings):
     WORKFLOW_LOG_CLEANUP_BATCH_SIZE: int = Field(
         default=100, description="Batch size for workflow run log cleanup operations"
     )
+    WORKFLOW_LOG_CLEANUP_SPECIFIC_WORKFLOW_IDS: str = Field(
+        default="", description="Comma-separated list of workflow IDs to clean logs for"
+    )
 
 
 class SwaggerUIConfig(BaseSettings):
@@ -1344,6 +1347,10 @@ class SandboxExpiredRecordsCleanConfig(BaseSettings):
         description="Maximum number of records to process in each batch",
         default=1000,
     )
+    SANDBOX_EXPIRED_RECORDS_CLEAN_BATCH_MAX_INTERVAL: PositiveInt = Field(
+        description="Maximum interval in milliseconds between batches",
+        default=200,
+    )
     SANDBOX_EXPIRED_RECORDS_RETENTION_DAYS: PositiveInt = Field(
         description="Retention days for sandbox expired workflow_run records and message records",
         default=30,

api/controllers/console/remote_files.py

@@ -1,6 +1,7 @@
 import urllib.parse
 
 import httpx
+from flask_restx import Resource
 from pydantic import BaseModel, Field
 
 import services
@@ -10,12 +11,12 @@ from controllers.common.errors import (
     RemoteFileUploadError,
     UnsupportedFileTypeError,
 )
-from controllers.fastopenapi import console_router
+from controllers.console import console_ns
 from core.file import helpers as file_helpers
 from core.helper import ssrf_proxy
 from extensions.ext_database import db
 from fields.file_fields import FileWithSignedUrl, RemoteFileInfo
-from libs.login import current_account_with_tenant
+from libs.login import current_account_with_tenant, login_required
 from services.file_service import FileService
 
 
@@ -23,69 +24,73 @@ class RemoteFileUploadPayload(BaseModel):
     url: str = Field(..., description="URL to fetch")
 
 
-@console_router.get(
-    "/remote-files/<path:url>",
-    response_model=RemoteFileInfo,
-    tags=["console"],
-)
-def get_remote_file_info(url: str) -> RemoteFileInfo:
-    decoded_url = urllib.parse.unquote(url)
-    resp = ssrf_proxy.head(decoded_url)
-    if resp.status_code != httpx.codes.OK:
-        resp = ssrf_proxy.get(decoded_url, timeout=3)
-    resp.raise_for_status()
-    return RemoteFileInfo(
-        file_type=resp.headers.get("Content-Type", "application/octet-stream"),
-        file_length=int(resp.headers.get("Content-Length", 0)),
-    )
-
-
-@console_router.post(
-    "/remote-files/upload",
-    response_model=FileWithSignedUrl,
-    tags=["console"],
-    status_code=201,
-)
-def upload_remote_file(payload: RemoteFileUploadPayload) -> FileWithSignedUrl:
-    url = payload.url
-
-    try:
-        resp = ssrf_proxy.head(url=url)
+@console_ns.route("/remote-files/<path:url>")
+class GetRemoteFileInfo(Resource):
+    @login_required
+    def get(self, url: str):
+        decoded_url = urllib.parse.unquote(url)
+        resp = ssrf_proxy.head(decoded_url)
         if resp.status_code != httpx.codes.OK:
-            resp = ssrf_proxy.get(url=url, timeout=3, follow_redirects=True)
-        if resp.status_code != httpx.codes.OK:
-            raise RemoteFileUploadError(f"Failed to fetch file from {url}: {resp.text}")
-    except httpx.RequestError as e:
-        raise RemoteFileUploadError(f"Failed to fetch file from {url}: {str(e)}")
+            resp = ssrf_proxy.get(decoded_url, timeout=3)
+        resp.raise_for_status()
+        return RemoteFileInfo(
+            file_type=resp.headers.get("Content-Type", "application/octet-stream"),
+            file_length=int(resp.headers.get("Content-Length", 0)),
+        ).model_dump(mode="json")
 
-    file_info = helpers.guess_file_info_from_response(resp)
 
-    if not FileService.is_file_size_within_limit(extension=file_info.extension, file_size=file_info.size):
-        raise FileTooLargeError
+@console_ns.route("/remote-files/upload")
+class RemoteFileUpload(Resource):
+    @login_required
+    def post(self):
+        payload = RemoteFileUploadPayload.model_validate(console_ns.payload)
+        url = payload.url
 
-    content = resp.content if resp.request.method == "GET" else ssrf_proxy.get(url).content
+        # Try to fetch remote file metadata/content first
+        try:
+            resp = ssrf_proxy.head(url=url)
+            if resp.status_code != httpx.codes.OK:
+                resp = ssrf_proxy.get(url=url, timeout=3, follow_redirects=True)
+            if resp.status_code != httpx.codes.OK:
+                # Normalize into a user-friendly error message expected by tests
+                raise RemoteFileUploadError(f"Failed to fetch file from {url}: {resp.text}")
+        except httpx.RequestError as e:
+            raise RemoteFileUploadError(f"Failed to fetch file from {url}: {str(e)}")
 
-    try:
-        user, _ = current_account_with_tenant()
-        upload_file = FileService(db.engine).upload_file(
-            filename=file_info.filename,
-            content=content,
-            mimetype=file_info.mimetype,
-            user=user,
-            source_url=url,
+        file_info = helpers.guess_file_info_from_response(resp)
+
+        # Enforce file size limit with 400 (Bad Request) per tests' expectation
+        if not FileService.is_file_size_within_limit(extension=file_info.extension, file_size=file_info.size):
+            raise FileTooLargeError()
+
+        # Load content if needed
+        content = resp.content if resp.request.method == "GET" else ssrf_proxy.get(url).content
+
+        try:
+            user, _ = current_account_with_tenant()
+            upload_file = FileService(db.engine).upload_file(
+                filename=file_info.filename,
+                content=content,
+                mimetype=file_info.mimetype,
+                user=user,
+                source_url=url,
+            )
+        except services.errors.file.FileTooLargeError as file_too_large_error:
+            raise FileTooLargeError(file_too_large_error.description)
+        except services.errors.file.UnsupportedFileTypeError:
+            raise UnsupportedFileTypeError()
+
+        # Success: return created resource with 201 status
+        return (
+            FileWithSignedUrl(
+                id=upload_file.id,
+                name=upload_file.name,
+                size=upload_file.size,
+                extension=upload_file.extension,
+                url=file_helpers.get_signed_file_url(upload_file_id=upload_file.id),
+                mime_type=upload_file.mime_type,
+                created_by=upload_file.created_by,
+                created_at=int(upload_file.created_at.timestamp()),
+            ).model_dump(mode="json"),
+            201,
         )
-    except services.errors.file.FileTooLargeError as file_too_large_error:
-        raise FileTooLargeError(file_too_large_error.description)
-    except services.errors.file.UnsupportedFileTypeError:
-        raise UnsupportedFileTypeError()
-
-    return FileWithSignedUrl(
-        id=upload_file.id,
-        name=upload_file.name,
-        size=upload_file.size,
-        extension=upload_file.extension,
-        url=file_helpers.get_signed_file_url(upload_file_id=upload_file.id),
-        mime_type=upload_file.mime_type,
-        created_by=upload_file.created_by,
-        created_at=int(upload_file.created_at.timestamp()),
-    )

api/migrations/versions/2026_02_11_1549-fce013ca180e_fix_index_to_optimize_message_clean_job_.py

@@ -0,0 +1,39 @@
+"""fix index to optimize message clean job performance
+
+Revision ID: fce013ca180e
+Revises: f55813ffe2c8
+Create Date: 2026-02-11 15:49:17.603638
+
+"""
+from alembic import op
+import models as models
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision = 'fce013ca180e'
+down_revision = 'f55813ffe2c8'
+branch_labels = None
+depends_on = None
+
+
+def upgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table('messages', schema=None) as batch_op:
+        batch_op.drop_index(batch_op.f('message_created_at_idx'))
+
+    with op.batch_alter_table('saved_messages', schema=None) as batch_op:
+        batch_op.create_index('saved_message_message_id_idx', ['message_id'], unique=False)
+
+    # ### end Alembic commands ###
+
+
+def downgrade():
+    # ### commands auto generated by Alembic - please adjust! ###
+    with op.batch_alter_table('saved_messages', schema=None) as batch_op:
+        batch_op.drop_index('saved_message_message_id_idx')
+
+    with op.batch_alter_table('messages', schema=None) as batch_op:
+        batch_op.create_index(batch_op.f('message_created_at_idx'), ['created_at'], unique=False)
+
+    # ### end Alembic commands ###

api/models/model.py

@@ -1040,7 +1040,6 @@ class Message(Base):
         Index("message_end_user_idx", "app_id", "from_source", "from_end_user_id"),
         Index("message_account_idx", "app_id", "from_source", "from_account_id"),
         Index("message_workflow_run_id_idx", "conversation_id", "workflow_run_id"),
-        Index("message_created_at_idx", "created_at"),
         Index("message_app_mode_idx", "app_mode"),
         Index("message_created_at_id_idx", "created_at", "id"),
     )

api/models/web.py

@@ -16,6 +16,7 @@ class SavedMessage(TypeBase):
     __table_args__ = (
         sa.PrimaryKeyConstraint("id", name="saved_message_pkey"),
         sa.Index("saved_message_message_idx", "app_id", "message_id", "created_by_role", "created_by"),
+        sa.Index("saved_message_message_id_idx", "message_id"),
     )
 
     id: Mapped[str] = mapped_column(

api/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "dify-api"
-version = "1.12.1"
+version = "1.13.0"
 requires-python = ">=3.11,<3.13"
 
 dependencies = [

api/repositories/api_workflow_run_repository.py

@@ -264,9 +264,15 @@ class APIWorkflowRunRepository(WorkflowExecutionRepository, Protocol):
         batch_size: int,
         run_types: Sequence[WorkflowType] | None = None,
         tenant_ids: Sequence[str] | None = None,
+        workflow_ids: Sequence[str] | None = None,
     ) -> Sequence[WorkflowRun]:
         """
         Fetch ended workflow runs in a time window for archival and clean batching.
+
+        Optional filters:
+        - run_types
+        - tenant_ids
+        - workflow_ids
         """
         ...
 

api/repositories/sqlalchemy_api_workflow_run_repository.py

@@ -386,6 +386,7 @@ class DifyAPISQLAlchemyWorkflowRunRepository(APIWorkflowRunRepository):
         batch_size: int,
         run_types: Sequence[WorkflowType] | None = None,
         tenant_ids: Sequence[str] | None = None,
+        workflow_ids: Sequence[str] | None = None,
     ) -> Sequence[WorkflowRun]:
         """
         Fetch ended workflow runs in a time window for archival and clean batching.
@@ -394,7 +395,7 @@ class DifyAPISQLAlchemyWorkflowRunRepository(APIWorkflowRunRepository):
         - created_at in [start_from, end_before)
         - type in run_types (when provided)
         - status is an ended state
-        - optional tenant_id filter and cursor (last_seen) for pagination
+        - optional tenant_id, workflow_id filters and cursor (last_seen) for pagination
         """
         with self._session_maker() as session:
             stmt = (
@@ -417,6 +418,9 @@ class DifyAPISQLAlchemyWorkflowRunRepository(APIWorkflowRunRepository):
             if tenant_ids:
                 stmt = stmt.where(WorkflowRun.tenant_id.in_(tenant_ids))
 
+            if workflow_ids:
+                stmt = stmt.where(WorkflowRun.workflow_id.in_(workflow_ids))
+
             if last_seen:
                 stmt = stmt.where(
                     or_(

api/schedule/clean_workflow_runlogs_precise.py

@@ -4,7 +4,6 @@ import time
 from collections.abc import Sequence
 
 import click
-from sqlalchemy import select
 from sqlalchemy.orm import Session, sessionmaker
 
 import app
@@ -13,6 +12,7 @@ from extensions.ext_database import db
 from models.model import (
     AppAnnotationHitHistory,
     Conversation,
+    DatasetRetrieverResource,
     Message,
     MessageAgentThought,
     MessageAnnotation,
@@ -20,7 +20,10 @@ from models.model import (
     MessageFeedback,
     MessageFile,
 )
-from models.workflow import ConversationVariable, WorkflowAppLog, WorkflowNodeExecutionModel, WorkflowRun
+from models.web import SavedMessage
+from models.workflow import ConversationVariable, WorkflowRun
+from repositories.factory import DifyAPIRepositoryFactory
+from repositories.sqlalchemy_workflow_trigger_log_repository import SQLAlchemyWorkflowTriggerLogRepository
 
 logger = logging.getLogger(__name__)
 
@@ -29,8 +32,15 @@ MAX_RETRIES = 3
 BATCH_SIZE = dify_config.WORKFLOW_LOG_CLEANUP_BATCH_SIZE
 
 
-@app.celery.task(queue="dataset")
-def clean_workflow_runlogs_precise():
+def _get_specific_workflow_ids() -> list[str]:
+    workflow_ids_str = dify_config.WORKFLOW_LOG_CLEANUP_SPECIFIC_WORKFLOW_IDS.strip()
+    if not workflow_ids_str:
+        return []
+    return [wid.strip() for wid in workflow_ids_str.split(",") if wid.strip()]
+
+
+@app.celery.task(queue="retention")
+def clean_workflow_runlogs_precise() -> None:
     """Clean expired workflow run logs with retry mechanism and complete message cascade"""
 
     click.echo(click.style("Start clean workflow run logs (precise mode with complete cascade).", fg="green"))
@@ -39,48 +49,48 @@ def clean_workflow_runlogs_precise():
     retention_days = dify_config.WORKFLOW_LOG_RETENTION_DAYS
     cutoff_date = datetime.datetime.now() - datetime.timedelta(days=retention_days)
     session_factory = sessionmaker(db.engine, expire_on_commit=False)
+    workflow_run_repo = DifyAPIRepositoryFactory.create_api_workflow_run_repository(session_factory)
+    workflow_ids = _get_specific_workflow_ids()
+    workflow_ids_filter = workflow_ids or None
 
     try:
-        with session_factory.begin() as session:
-            total_workflow_runs = session.query(WorkflowRun).where(WorkflowRun.created_at < cutoff_date).count()
-            if total_workflow_runs == 0:
-                logger.info("No expired workflow run logs found")
-                return
-            logger.info("Found %s expired workflow run logs to clean", total_workflow_runs)
-
         total_deleted = 0
         failed_batches = 0
         batch_count = 0
+        last_seen: tuple[datetime.datetime, str] | None = None
         while True:
+            run_rows = workflow_run_repo.get_runs_batch_by_time_range(
+                start_from=None,
+                end_before=cutoff_date,
+                last_seen=last_seen,
+                batch_size=BATCH_SIZE,
+                workflow_ids=workflow_ids_filter,
+            )
+
+            if not run_rows:
+                if batch_count == 0:
+                    logger.info("No expired workflow run logs found")
+                break
+
+            last_seen = (run_rows[-1].created_at, run_rows[-1].id)
+            batch_count += 1
             with session_factory.begin() as session:
-                workflow_run_ids = session.scalars(
-                    select(WorkflowRun.id)
-                    .where(WorkflowRun.created_at < cutoff_date)
-                    .order_by(WorkflowRun.created_at, WorkflowRun.id)
-                    .limit(BATCH_SIZE)
-                ).all()
+                success = _delete_batch(session, workflow_run_repo, run_rows, failed_batches)
 
-                if not workflow_run_ids:
+            if success:
+                total_deleted += len(run_rows)
+                failed_batches = 0
+            else:
+                failed_batches += 1
+                if failed_batches >= MAX_RETRIES:
+                    logger.error("Failed to delete batch after %s retries, aborting cleanup for today", MAX_RETRIES)
                     break
-
-                batch_count += 1
-
-                success = _delete_batch(session, workflow_run_ids, failed_batches)
-
-                if success:
-                    total_deleted += len(workflow_run_ids)
-                    failed_batches = 0
                 else:
-                    failed_batches += 1
-                    if failed_batches >= MAX_RETRIES:
-                        logger.error("Failed to delete batch after %s retries, aborting cleanup for today", MAX_RETRIES)
-                        break
-                    else:
-                        # Calculate incremental delay times: 5, 10, 15 minutes
-                        retry_delay_minutes = failed_batches * 5
-                        logger.warning("Batch deletion failed, retrying in %s minutes...", retry_delay_minutes)
-                        time.sleep(retry_delay_minutes * 60)
-                        continue
+                    # Calculate incremental delay times: 5, 10, 15 minutes
+                    retry_delay_minutes = failed_batches * 5
+                    logger.warning("Batch deletion failed, retrying in %s minutes...", retry_delay_minutes)
+                    time.sleep(retry_delay_minutes * 60)
+                    continue
 
         logger.info("Cleanup completed: %s expired workflow run logs deleted", total_deleted)
 
@@ -93,10 +103,16 @@ def clean_workflow_runlogs_precise():
     click.echo(click.style(f"Cleaned workflow run logs from db success latency: {execution_time:.2f}s", fg="green"))
 
 
-def _delete_batch(session: Session, workflow_run_ids: Sequence[str], attempt_count: int) -> bool:
+def _delete_batch(
+    session: Session,
+    workflow_run_repo,
+    workflow_runs: Sequence[WorkflowRun],
+    attempt_count: int,
+) -> bool:
     """Delete a single batch of workflow runs and all related data within a nested transaction."""
     try:
         with session.begin_nested():
+            workflow_run_ids = [run.id for run in workflow_runs]
             message_data = (
                 session.query(Message.id, Message.conversation_id)
                 .where(Message.workflow_run_id.in_(workflow_run_ids))
@@ -107,11 +123,13 @@ def _delete_batch(session: Session, workflow_run_ids: Sequence[str], attempt_cou
             if message_id_list:
                 message_related_models = [
                     AppAnnotationHitHistory,
+                    DatasetRetrieverResource,
                     MessageAgentThought,
                     MessageChain,
                     MessageFile,
                     MessageAnnotation,
                     MessageFeedback,
+                    SavedMessage,
                 ]
                 for model in message_related_models:
                     session.query(model).where(model.message_id.in_(message_id_list)).delete(synchronize_session=False)  # type: ignore
@@ -122,14 +140,6 @@ def _delete_batch(session: Session, workflow_run_ids: Sequence[str], attempt_cou
                     synchronize_session=False
                 )
 
-            session.query(WorkflowAppLog).where(WorkflowAppLog.workflow_run_id.in_(workflow_run_ids)).delete(
-                synchronize_session=False
-            )
-
-            session.query(WorkflowNodeExecutionModel).where(
-                WorkflowNodeExecutionModel.workflow_run_id.in_(workflow_run_ids)
-            ).delete(synchronize_session=False)
-
             if conversation_id_list:
                 session.query(ConversationVariable).where(
                     ConversationVariable.conversation_id.in_(conversation_id_list)
@@ -139,7 +149,22 @@ def _delete_batch(session: Session, workflow_run_ids: Sequence[str], attempt_cou
                     synchronize_session=False
                 )
 
-            session.query(WorkflowRun).where(WorkflowRun.id.in_(workflow_run_ids)).delete(synchronize_session=False)
+            def _delete_node_executions(active_session: Session, runs: Sequence[WorkflowRun]) -> tuple[int, int]:
+                run_ids = [run.id for run in runs]
+                repo = DifyAPIRepositoryFactory.create_api_workflow_node_execution_repository(
+                    session_maker=sessionmaker(bind=active_session.get_bind(), expire_on_commit=False)
+                )
+                return repo.delete_by_runs(active_session, run_ids)
+
+            def _delete_trigger_logs(active_session: Session, run_ids: Sequence[str]) -> int:
+                trigger_repo = SQLAlchemyWorkflowTriggerLogRepository(active_session)
+                return trigger_repo.delete_by_run_ids(run_ids)
+
+            workflow_run_repo.delete_runs_with_related(
+                workflow_runs,
+                delete_node_executions=_delete_node_executions,
+                delete_trigger_logs=_delete_trigger_logs,
+            )
 
             return True
 

api/services/retention/conversation/messages_clean_service.py

@@ -1,10 +1,13 @@
 import datetime
 import logging
+import os
 import random
+import time
 from collections.abc import Sequence
 from typing import cast
 
-from sqlalchemy import delete, select
+import sqlalchemy as sa
+from sqlalchemy import delete, select, tuple_
 from sqlalchemy.engine import CursorResult
 from sqlalchemy.orm import Session
 
@@ -193,11 +196,15 @@ class MessagesCleanService:
             self._end_before,
         )
 
+        max_batch_interval_ms = int(os.environ.get("SANDBOX_EXPIRED_RECORDS_CLEAN_BATCH_MAX_INTERVAL", 200))
+
         while True:
             stats["batches"] += 1
+            batch_start = time.monotonic()
 
             # Step 1: Fetch a batch of messages using cursor
             with Session(db.engine, expire_on_commit=False) as session:
+                fetch_messages_start = time.monotonic()
                 msg_stmt = (
                     select(Message.id, Message.app_id, Message.created_at)
                     .where(Message.created_at < self._end_before)
@@ -209,13 +216,13 @@ class MessagesCleanService:
                     msg_stmt = msg_stmt.where(Message.created_at >= self._start_from)
 
                 # Apply cursor condition: (created_at, id) > (last_created_at, last_message_id)
-                # This translates to:
-                #   created_at > last_created_at OR (created_at = last_created_at AND id > last_message_id)
                 if _cursor:
-                    # Continuing from previous batch
                     msg_stmt = msg_stmt.where(
-                        (Message.created_at > _cursor[0])
-                        | ((Message.created_at == _cursor[0]) & (Message.id > _cursor[1]))
+                        tuple_(Message.created_at, Message.id)
+                        > tuple_(
+                            sa.literal(_cursor[0], type_=sa.DateTime()),
+                            sa.literal(_cursor[1], type_=Message.id.type),
+                        )
                     )
 
                 raw_messages = list(session.execute(msg_stmt).all())
@@ -223,6 +230,12 @@ class MessagesCleanService:
                     SimpleMessage(id=msg_id, app_id=app_id, created_at=msg_created_at)
                     for msg_id, app_id, msg_created_at in raw_messages
                 ]
+                logger.info(
+                    "clean_messages (batch %s): fetched %s messages in %sms",
+                    stats["batches"],
+                    len(messages),
+                    int((time.monotonic() - fetch_messages_start) * 1000),
+                )
 
                 # Track total messages fetched across all batches
                 stats["total_messages"] += len(messages)
@@ -241,8 +254,16 @@ class MessagesCleanService:
                     logger.info("clean_messages (batch %s): no app_ids found, skip", stats["batches"])
                     continue
 
+                fetch_apps_start = time.monotonic()
                 app_stmt = select(App.id, App.tenant_id).where(App.id.in_(app_ids))
                 apps = list(session.execute(app_stmt).all())
+                logger.info(
+                    "clean_messages (batch %s): fetched %s apps for %s app_ids in %sms",
+                    stats["batches"],
+                    len(apps),
+                    len(app_ids),
+                    int((time.monotonic() - fetch_apps_start) * 1000),
+                )
 
             if not apps:
                 logger.info("clean_messages (batch %s): no apps found, skip", stats["batches"])
@@ -252,7 +273,15 @@ class MessagesCleanService:
             app_to_tenant: dict[str, str] = {app.id: app.tenant_id for app in apps}
 
             # Step 3: Delegate to policy to determine which messages to delete
+            policy_start = time.monotonic()
             message_ids_to_delete = self._policy.filter_message_ids(messages, app_to_tenant)
+            logger.info(
+                "clean_messages (batch %s): policy selected %s/%s messages in %sms",
+                stats["batches"],
+                len(message_ids_to_delete),
+                len(messages),
+                int((time.monotonic() - policy_start) * 1000),
+            )
 
             if not message_ids_to_delete:
                 logger.info("clean_messages (batch %s): no messages to delete, skip", stats["batches"])
@@ -263,14 +292,20 @@ class MessagesCleanService:
             # Step 4: Batch delete messages and their relations
             if not self._dry_run:
                 with Session(db.engine, expire_on_commit=False) as session:
+                    delete_relations_start = time.monotonic()
                     # Delete related records first
                     self._batch_delete_message_relations(session, message_ids_to_delete)
+                    delete_relations_ms = int((time.monotonic() - delete_relations_start) * 1000)
 
                     # Delete messages
+                    delete_messages_start = time.monotonic()
                     delete_stmt = delete(Message).where(Message.id.in_(message_ids_to_delete))
                     delete_result = cast(CursorResult, session.execute(delete_stmt))
                     messages_deleted = delete_result.rowcount
+                    delete_messages_ms = int((time.monotonic() - delete_messages_start) * 1000)
+                    commit_start = time.monotonic()
                     session.commit()
+                    commit_ms = int((time.monotonic() - commit_start) * 1000)
 
                     stats["total_deleted"] += messages_deleted
 
@@ -280,6 +315,19 @@ class MessagesCleanService:
                         len(messages),
                         messages_deleted,
                     )
+                    logger.info(
+                        "clean_messages (batch %s): relations %sms,  messages %sms, commit %sms, batch total %sms",
+                        stats["batches"],
+                        delete_relations_ms,
+                        delete_messages_ms,
+                        commit_ms,
+                        int((time.monotonic() - batch_start) * 1000),
+                    )
+
+                # Random sleep between batches to avoid overwhelming the database
+                sleep_ms = random.uniform(0, max_batch_interval_ms)  # noqa: S311
+                logger.info("clean_messages (batch %s): sleeping for %.2fms", stats["batches"], sleep_ms)
+                time.sleep(sleep_ms / 1000)
             else:
                 # Log random sample of message IDs that would be deleted (up to 10)
                 sample_size = min(10, len(message_ids_to_delete))

api/services/retention/workflow_run/clear_free_plan_expired_workflow_run_logs.py

@@ -1,5 +1,8 @@
 import datetime
 import logging
+import os
+import random
+import time
 from collections.abc import Iterable, Sequence
 
 import click
@@ -72,7 +75,12 @@ class WorkflowRunCleanup:
         batch_index = 0
         last_seen: tuple[datetime.datetime, str] | None = None
 
+        max_batch_interval_ms = int(os.environ.get("SANDBOX_EXPIRED_RECORDS_CLEAN_BATCH_MAX_INTERVAL", 200))
+
         while True:
+            batch_start = time.monotonic()
+
+            fetch_start = time.monotonic()
             run_rows = self.workflow_run_repo.get_runs_batch_by_time_range(
                 start_from=self.window_start,
                 end_before=self.window_end,
@@ -80,12 +88,30 @@ class WorkflowRunCleanup:
                 batch_size=self.batch_size,
             )
             if not run_rows:
+                logger.info("workflow_run_cleanup (batch #%s): no more rows to process", batch_index + 1)
                 break
 
             batch_index += 1
             last_seen = (run_rows[-1].created_at, run_rows[-1].id)
+            logger.info(
+                "workflow_run_cleanup (batch #%s): fetched %s rows in %sms",
+                batch_index,
+                len(run_rows),
+                int((time.monotonic() - fetch_start) * 1000),
+            )
+
             tenant_ids = {row.tenant_id for row in run_rows}
+
+            filter_start = time.monotonic()
             free_tenants = self._filter_free_tenants(tenant_ids)
+            logger.info(
+                "workflow_run_cleanup (batch #%s): filtered %s free tenants from %s tenants in %sms",
+                batch_index,
+                len(free_tenants),
+                len(tenant_ids),
+                int((time.monotonic() - filter_start) * 1000),
+            )
+
             free_runs = [row for row in run_rows if row.tenant_id in free_tenants]
             paid_or_skipped = len(run_rows) - len(free_runs)
 
@@ -104,11 +130,17 @@ class WorkflowRunCleanup:
             total_runs_targeted += len(free_runs)
 
             if self.dry_run:
+                count_start = time.monotonic()
                 batch_counts = self.workflow_run_repo.count_runs_with_related(
                     free_runs,
                     count_node_executions=self._count_node_executions,
                     count_trigger_logs=self._count_trigger_logs,
                 )
+                logger.info(
+                    "workflow_run_cleanup (batch #%s, dry_run): counted related records in %sms",
+                    batch_index,
+                    int((time.monotonic() - count_start) * 1000),
+                )
                 if related_totals is not None:
                     for key in related_totals:
                         related_totals[key] += batch_counts.get(key, 0)
@@ -120,14 +152,21 @@ class WorkflowRunCleanup:
                         fg="yellow",
                     )
                 )
+                logger.info(
+                    "workflow_run_cleanup (batch #%s, dry_run): batch total %sms",
+                    batch_index,
+                    int((time.monotonic() - batch_start) * 1000),
+                )
                 continue
 
             try:
+                delete_start = time.monotonic()
                 counts = self.workflow_run_repo.delete_runs_with_related(
                     free_runs,
                     delete_node_executions=self._delete_node_executions,
                     delete_trigger_logs=self._delete_trigger_logs,
                 )
+                delete_ms = int((time.monotonic() - delete_start) * 1000)
             except Exception:
                 logger.exception("Failed to delete workflow runs batch ending at %s", last_seen[0])
                 raise
@@ -143,6 +182,17 @@ class WorkflowRunCleanup:
                     fg="green",
                 )
             )
+            logger.info(
+                "workflow_run_cleanup (batch #%s): delete %sms, batch total %sms",
+                batch_index,
+                delete_ms,
+                int((time.monotonic() - batch_start) * 1000),
+            )
+
+            # Random sleep between batches to avoid overwhelming the database
+            sleep_ms = random.uniform(0, max_batch_interval_ms)  # noqa: S311
+            logger.info("workflow_run_cleanup (batch #%s): sleeping for %.2fms", batch_index, sleep_ms)
+            time.sleep(sleep_ms / 1000)
 
         if self.dry_run:
             if self.window_start:

api/tests/unit_tests/controllers/console/test_fastopenapi_remote_files.py

@@ -1,92 +1,286 @@
-import builtins
+"""Tests for remote file upload API endpoints using Flask-RESTX."""
+
+import contextlib
 from datetime import datetime
 from types import SimpleNamespace
-from unittest.mock import patch
+from unittest.mock import Mock, patch
 
 import httpx
 import pytest
-from flask import Flask
-from flask.views import MethodView
-
-from extensions import ext_fastopenapi
-
-if not hasattr(builtins, "MethodView"):
-    builtins.MethodView = MethodView  # type: ignore[attr-defined]
+from flask import Flask, g
 
 
 @pytest.fixture
 def app() -> Flask:
+    """Create Flask app for testing."""
     app = Flask(__name__)
     app.config["TESTING"] = True
+    app.config["SECRET_KEY"] = "test-secret-key"
     return app
 
 
-def test_console_remote_files_fastopenapi_get_info(app: Flask):
-    ext_fastopenapi.init_app(app)
+@pytest.fixture
+def client(app):
+    """Create test client with console blueprint registered."""
+    from controllers.console import bp
 
-    response = httpx.Response(
-        200,
-        request=httpx.Request("HEAD", "http://example.com/file.txt"),
-        headers={"Content-Type": "text/plain", "Content-Length": "10"},
-    )
-
-    with patch("controllers.console.remote_files.ssrf_proxy.head", return_value=response):
-        client = app.test_client()
-        encoded_url = "http%3A%2F%2Fexample.com%2Ffile.txt"
-        resp = client.get(f"/console/api/remote-files/{encoded_url}")
-
-    assert resp.status_code == 200
-    assert resp.get_json() == {"file_type": "text/plain", "file_length": 10}
+    app.register_blueprint(bp)
+    return app.test_client()
 
 
-def test_console_remote_files_fastopenapi_upload(app: Flask):
-    ext_fastopenapi.init_app(app)
+@pytest.fixture
+def mock_account():
+    """Create a mock account for testing."""
+    from models import Account
 
-    head_response = httpx.Response(
-        200,
-        request=httpx.Request("GET", "http://example.com/file.txt"),
-        content=b"hello",
-    )
-    file_info = SimpleNamespace(
-        extension="txt",
-        size=5,
-        filename="file.txt",
-        mimetype="text/plain",
-    )
-    uploaded = SimpleNamespace(
-        id="file-id",
-        name="file.txt",
-        size=5,
-        extension="txt",
-        mime_type="text/plain",
-        created_by="user-id",
-        created_at=datetime(2024, 1, 1),
-    )
+    account = Mock(spec=Account)
+    account.id = "test-account-id"
+    account.current_tenant_id = "test-tenant-id"
+    return account
 
-    with (
-        patch("controllers.console.remote_files.db", new=SimpleNamespace(engine=object())),
-        patch("controllers.console.remote_files.ssrf_proxy.head", return_value=head_response),
-        patch("controllers.console.remote_files.helpers.guess_file_info_from_response", return_value=file_info),
-        patch("controllers.console.remote_files.FileService.is_file_size_within_limit", return_value=True),
-        patch("controllers.console.remote_files.FileService.__init__", return_value=None),
-        patch("controllers.console.remote_files.current_account_with_tenant", return_value=(object(), "tenant-id")),
-        patch("controllers.console.remote_files.FileService.upload_file", return_value=uploaded),
-        patch("controllers.console.remote_files.file_helpers.get_signed_file_url", return_value="signed-url"),
-    ):
-        client = app.test_client()
-        resp = client.post(
-            "/console/api/remote-files/upload",
-            json={"url": "http://example.com/file.txt"},
+
+@pytest.fixture
+def auth_ctx(app, mock_account):
+    """Context manager to set auth/tenant context in flask.g for a request."""
+
+    @contextlib.contextmanager
+    def _ctx():
+        with app.test_request_context():
+            g._login_user = mock_account
+            g._current_tenant = mock_account.current_tenant_id
+            yield
+
+    return _ctx
+
+
+class TestGetRemoteFileInfo:
+    """Test GET /console/api/remote-files/<path:url> endpoint."""
+
+    def test_get_remote_file_info_success(self, app, client, mock_account):
+        """Test successful retrieval of remote file info."""
+        response = httpx.Response(
+            200,
+            request=httpx.Request("HEAD", "http://example.com/file.txt"),
+            headers={"Content-Type": "text/plain", "Content-Length": "1024"},
         )
 
-    assert resp.status_code == 201
-    assert resp.get_json() == {
-        "id": "file-id",
-        "name": "file.txt",
-        "size": 5,
-        "extension": "txt",
-        "url": "signed-url",
-        "mime_type": "text/plain",
-        "created_by": "user-id",
-        "created_at": int(uploaded.created_at.timestamp()),
-    }
+        with (
+            patch(
+                "controllers.console.remote_files.current_account_with_tenant",
+                return_value=(mock_account, "test-tenant-id"),
+            ),
+            patch("controllers.console.remote_files.ssrf_proxy.head", return_value=response),
+            patch("libs.login.check_csrf_token", return_value=None),
+        ):
+            with app.test_request_context():
+                g._login_user = mock_account
+                g._current_tenant = mock_account.current_tenant_id
+                encoded_url = "http%3A%2F%2Fexample.com%2Ffile.txt"
+                resp = client.get(f"/console/api/remote-files/{encoded_url}")
+
+        assert resp.status_code == 200
+        data = resp.get_json()
+        assert data["file_type"] == "text/plain"
+        assert data["file_length"] == 1024
+
+    def test_get_remote_file_info_fallback_to_get_on_head_failure(self, app, client, mock_account):
+        """Test fallback to GET when HEAD returns non-200 status."""
+        head_response = httpx.Response(
+            404,
+            request=httpx.Request("HEAD", "http://example.com/file.pdf"),
+        )
+        get_response = httpx.Response(
+            200,
+            request=httpx.Request("GET", "http://example.com/file.pdf"),
+            headers={"Content-Type": "application/pdf", "Content-Length": "2048"},
+        )
+
+        with (
+            patch(
+                "controllers.console.remote_files.current_account_with_tenant",
+                return_value=(mock_account, "test-tenant-id"),
+            ),
+            patch("controllers.console.remote_files.ssrf_proxy.head", return_value=head_response),
+            patch("controllers.console.remote_files.ssrf_proxy.get", return_value=get_response),
+            patch("libs.login.check_csrf_token", return_value=None),
+        ):
+            with app.test_request_context():
+                g._login_user = mock_account
+                g._current_tenant = mock_account.current_tenant_id
+                encoded_url = "http%3A%2F%2Fexample.com%2Ffile.pdf"
+                resp = client.get(f"/console/api/remote-files/{encoded_url}")
+
+        assert resp.status_code == 200
+        data = resp.get_json()
+        assert data["file_type"] == "application/pdf"
+        assert data["file_length"] == 2048
+
+
+class TestRemoteFileUpload:
+    """Test POST /console/api/remote-files/upload endpoint."""
+
+    @pytest.mark.parametrize(
+        ("head_status", "use_get"),
+        [
+            (200, False),  # HEAD succeeds
+            (405, True),  # HEAD fails -> fallback GET
+        ],
+    )
+    def test_upload_remote_file_success_paths(self, client, mock_account, auth_ctx, head_status, use_get):
+        url = "http://example.com/file.pdf"
+        head_resp = httpx.Response(
+            head_status,
+            request=httpx.Request("HEAD", url),
+            headers={"Content-Type": "application/pdf", "Content-Length": "1024"},
+        )
+        get_resp = httpx.Response(
+            200,
+            request=httpx.Request("GET", url),
+            headers={"Content-Type": "application/pdf", "Content-Length": "1024"},
+            content=b"file content",
+        )
+
+        file_info = SimpleNamespace(
+            extension="pdf",
+            size=1024,
+            filename="file.pdf",
+            mimetype="application/pdf",
+        )
+        uploaded_file = SimpleNamespace(
+            id="uploaded-file-id",
+            name="file.pdf",
+            size=1024,
+            extension="pdf",
+            mime_type="application/pdf",
+            created_by="test-account-id",
+            created_at=datetime(2024, 1, 1, 12, 0, 0),
+        )
+
+        with (
+            patch(
+                "controllers.console.remote_files.current_account_with_tenant",
+                return_value=(mock_account, "test-tenant-id"),
+            ),
+            patch("controllers.console.remote_files.ssrf_proxy.head", return_value=head_resp) as p_head,
+            patch("controllers.console.remote_files.ssrf_proxy.get", return_value=get_resp) as p_get,
+            patch(
+                "controllers.console.remote_files.helpers.guess_file_info_from_response",
+                return_value=file_info,
+            ),
+            patch(
+                "controllers.console.remote_files.FileService.is_file_size_within_limit",
+                return_value=True,
+            ),
+            patch("controllers.console.remote_files.db", spec=["engine"]),
+            patch("controllers.console.remote_files.FileService") as mock_file_service,
+            patch(
+                "controllers.console.remote_files.file_helpers.get_signed_file_url",
+                return_value="http://example.com/signed-url",
+            ),
+            patch("libs.login.check_csrf_token", return_value=None),
+        ):
+            mock_file_service.return_value.upload_file.return_value = uploaded_file
+
+            with auth_ctx():
+                resp = client.post(
+                    "/console/api/remote-files/upload",
+                    json={"url": url},
+                )
+
+        assert resp.status_code == 201
+        p_head.assert_called_once()
+        # GET is used either for fallback (HEAD fails) or to fetch content after HEAD succeeds
+        p_get.assert_called_once()
+        mock_file_service.return_value.upload_file.assert_called_once()
+
+        data = resp.get_json()
+        assert data["id"] == "uploaded-file-id"
+        assert data["name"] == "file.pdf"
+        assert data["size"] == 1024
+        assert data["extension"] == "pdf"
+        assert data["url"] == "http://example.com/signed-url"
+        assert data["mime_type"] == "application/pdf"
+        assert data["created_by"] == "test-account-id"
+
+    @pytest.mark.parametrize(
+        ("size_ok", "raises", "expected_status", "expected_msg"),
+        [
+            # When size check fails in controller, API returns 413 with message "File size exceeded..."
+            (False, None, 413, "file size exceeded"),
+            # When service raises unsupported type, controller maps to 415 with message "File type not allowed."
+            (True, "unsupported", 415, "file type not allowed"),
+        ],
+    )
+    def test_upload_remote_file_errors(
+        self, client, mock_account, auth_ctx, size_ok, raises, expected_status, expected_msg
+    ):
+        url = "http://example.com/x.pdf"
+        head_resp = httpx.Response(
+            200,
+            request=httpx.Request("HEAD", url),
+            headers={"Content-Type": "application/pdf", "Content-Length": "9"},
+        )
+        file_info = SimpleNamespace(extension="pdf", size=9, filename="x.pdf", mimetype="application/pdf")
+
+        with (
+            patch(
+                "controllers.console.remote_files.current_account_with_tenant",
+                return_value=(mock_account, "test-tenant-id"),
+            ),
+            patch("controllers.console.remote_files.ssrf_proxy.head", return_value=head_resp),
+            patch(
+                "controllers.console.remote_files.helpers.guess_file_info_from_response",
+                return_value=file_info,
+            ),
+            patch(
+                "controllers.console.remote_files.FileService.is_file_size_within_limit",
+                return_value=size_ok,
+            ),
+            patch("controllers.console.remote_files.db", spec=["engine"]),
+            patch("libs.login.check_csrf_token", return_value=None),
+        ):
+            if raises == "unsupported":
+                from services.errors.file import UnsupportedFileTypeError
+
+                with patch("controllers.console.remote_files.FileService") as mock_file_service:
+                    mock_file_service.return_value.upload_file.side_effect = UnsupportedFileTypeError("bad")
+                    with auth_ctx():
+                        resp = client.post(
+                            "/console/api/remote-files/upload",
+                            json={"url": url},
+                        )
+            else:
+                with auth_ctx():
+                    resp = client.post(
+                        "/console/api/remote-files/upload",
+                        json={"url": url},
+                    )
+
+        assert resp.status_code == expected_status
+        data = resp.get_json()
+        msg = (data.get("error") or {}).get("message") or data.get("message", "")
+        assert expected_msg in msg.lower()
+
+    def test_upload_remote_file_fetch_failure(self, client, mock_account, auth_ctx):
+        """Test upload when fetching of remote file fails."""
+        with (
+            patch(
+                "controllers.console.remote_files.current_account_with_tenant",
+                return_value=(mock_account, "test-tenant-id"),
+            ),
+            patch(
+                "controllers.console.remote_files.ssrf_proxy.head",
+                side_effect=httpx.RequestError("Connection failed"),
+            ),
+            patch("libs.login.check_csrf_token", return_value=None),
+        ):
+            with auth_ctx():
+                resp = client.post(
+                    "/console/api/remote-files/upload",
+                    json={"url": "http://unreachable.com/file.pdf"},
+                )
+
+        assert resp.status_code == 400
+        data = resp.get_json()
+        msg = (data.get("error") or {}).get("message") or data.get("message", "")
+        assert "failed to fetch" in msg.lower()

api/tests/unit_tests/core/schemas/test_resolver.py

@@ -496,6 +496,9 @@ class TestSchemaResolverClass:
         avg_time_no_cache = sum(results1) / len(results1)
 
         # Second run (with cache) - run multiple times
+        # Warm up cache first
+        resolve_dify_schema_refs(schema)
+
         results2 = []
         for _ in range(3):
             start = time.perf_counter()

api/tests/unit_tests/services/test_clear_free_plan_expired_workflow_run_logs.py

@@ -62,6 +62,9 @@ class FakeRepo:
         end_before: datetime.datetime,
         last_seen: tuple[datetime.datetime, str] | None,
         batch_size: int,
+        run_types=None,
+        tenant_ids=None,
+        workflow_ids=None,
     ) -> list[FakeRun]:
         if self.call_idx >= len(self.batches):
             return []

api/uv.lock

@@ -1366,7 +1366,7 @@ wheels = [
 
 [[package]]
 name = "dify-api"
-version = "1.12.1"
+version = "1.13.0"
 source = { virtual = "." }
 dependencies = [
     { name = "aliyun-log-python-sdk" },

docker/.env.example

@@ -1073,6 +1073,8 @@ WORKFLOW_LOG_CLEANUP_ENABLED=false
 WORKFLOW_LOG_RETENTION_DAYS=30
 # Batch size for workflow log cleanup operations (default: 100)
 WORKFLOW_LOG_CLEANUP_BATCH_SIZE=100
+# Comma-separated list of workflow IDs to clean logs for
+WORKFLOW_LOG_CLEANUP_SPECIFIC_WORKFLOW_IDS=
 
 # Aliyun SLS Logstore Configuration
 # Aliyun Access Key ID
@@ -1523,6 +1525,7 @@ AMPLITUDE_API_KEY=
 # Sandbox expired records clean configuration
 SANDBOX_EXPIRED_RECORDS_CLEAN_GRACEFUL_PERIOD=21
 SANDBOX_EXPIRED_RECORDS_CLEAN_BATCH_SIZE=1000
+SANDBOX_EXPIRED_RECORDS_CLEAN_BATCH_MAX_INTERVAL=200
 SANDBOX_EXPIRED_RECORDS_RETENTION_DAYS=30
 
 

docker/docker-compose-template.yaml

@@ -21,7 +21,7 @@ services:
 
   # API service
   api:
-    image: langgenius/dify-api:1.12.1
+    image: langgenius/dify-api:1.13.0
     restart: always
     environment:
       # Use the shared environment variables.
@@ -63,7 +63,7 @@ services:
   # worker service
   # The Celery worker for processing all queues (dataset, workflow, mail, etc.)
   worker:
-    image: langgenius/dify-api:1.12.1
+    image: langgenius/dify-api:1.13.0
     restart: always
     environment:
       # Use the shared environment variables.
@@ -102,7 +102,7 @@ services:
   # worker_beat service
   # Celery beat for scheduling periodic tasks.
   worker_beat:
-    image: langgenius/dify-api:1.12.1
+    image: langgenius/dify-api:1.13.0
     restart: always
     environment:
       # Use the shared environment variables.
@@ -132,7 +132,7 @@ services:
 
   # Frontend web application.
   web:
-    image: langgenius/dify-web:1.12.1
+    image: langgenius/dify-web:1.13.0
     restart: always
     environment:
       CONSOLE_API_URL: ${CONSOLE_API_URL:-}

docker/docker-compose.yaml

@@ -470,6 +470,7 @@ x-shared-env: &shared-api-worker-env
   WORKFLOW_LOG_CLEANUP_ENABLED: ${WORKFLOW_LOG_CLEANUP_ENABLED:-false}
   WORKFLOW_LOG_RETENTION_DAYS: ${WORKFLOW_LOG_RETENTION_DAYS:-30}
   WORKFLOW_LOG_CLEANUP_BATCH_SIZE: ${WORKFLOW_LOG_CLEANUP_BATCH_SIZE:-100}
+  WORKFLOW_LOG_CLEANUP_SPECIFIC_WORKFLOW_IDS: ${WORKFLOW_LOG_CLEANUP_SPECIFIC_WORKFLOW_IDS:-}
   ALIYUN_SLS_ACCESS_KEY_ID: ${ALIYUN_SLS_ACCESS_KEY_ID:-}
   ALIYUN_SLS_ACCESS_KEY_SECRET: ${ALIYUN_SLS_ACCESS_KEY_SECRET:-}
   ALIYUN_SLS_ENDPOINT: ${ALIYUN_SLS_ENDPOINT:-}
@@ -684,6 +685,7 @@ x-shared-env: &shared-api-worker-env
   AMPLITUDE_API_KEY: ${AMPLITUDE_API_KEY:-}
   SANDBOX_EXPIRED_RECORDS_CLEAN_GRACEFUL_PERIOD: ${SANDBOX_EXPIRED_RECORDS_CLEAN_GRACEFUL_PERIOD:-21}
   SANDBOX_EXPIRED_RECORDS_CLEAN_BATCH_SIZE: ${SANDBOX_EXPIRED_RECORDS_CLEAN_BATCH_SIZE:-1000}
+  SANDBOX_EXPIRED_RECORDS_CLEAN_BATCH_MAX_INTERVAL: ${SANDBOX_EXPIRED_RECORDS_CLEAN_BATCH_MAX_INTERVAL:-200}
   SANDBOX_EXPIRED_RECORDS_RETENTION_DAYS: ${SANDBOX_EXPIRED_RECORDS_RETENTION_DAYS:-30}
   PUBSUB_REDIS_URL: ${PUBSUB_REDIS_URL:-}
   PUBSUB_REDIS_CHANNEL_TYPE: ${PUBSUB_REDIS_CHANNEL_TYPE:-pubsub}
@@ -714,7 +716,7 @@ services:
 
   # API service
   api:
-    image: langgenius/dify-api:1.12.1
+    image: langgenius/dify-api:1.13.0
     restart: always
     environment:
       # Use the shared environment variables.
@@ -756,7 +758,7 @@ services:
   # worker service
   # The Celery worker for processing all queues (dataset, workflow, mail, etc.)
   worker:
-    image: langgenius/dify-api:1.12.1
+    image: langgenius/dify-api:1.13.0
     restart: always
     environment:
       # Use the shared environment variables.
@@ -795,7 +797,7 @@ services:
   # worker_beat service
   # Celery beat for scheduling periodic tasks.
   worker_beat:
-    image: langgenius/dify-api:1.12.1
+    image: langgenius/dify-api:1.13.0
     restart: always
     environment:
       # Use the shared environment variables.
@@ -825,7 +827,7 @@ services:
 
   # Frontend web application.
   web:
-    image: langgenius/dify-web:1.12.1
+    image: langgenius/dify-web:1.13.0
     restart: always
     environment:
       CONSOLE_API_URL: ${CONSOLE_API_URL:-}

web/app/components/header/account-setting/members-page/index.tsx

@@ -104,7 +104,7 @@ const MembersPage = () => {
             <UpgradeBtn className="mr-2" loc="member-invite" />
           )}
           <div className="shrink-0">
-            <InviteButton disabled={!isCurrentWorkspaceManager || isMemberFull} onClick={() => setInviteModalVisible(true)} />
+            {isCurrentWorkspaceManager && <InviteButton disabled={isMemberFull} onClick={() => setInviteModalVisible(true)} />}
           </div>
         </div>
         <div className="overflow-visible lg:overflow-visible">

web/app/components/rag-pipeline/hooks/use-pipeline-run.spec.ts

@@ -92,8 +92,10 @@ vi.mock('@/service/workflow', () => ({
 }))
 
 const mockInvalidAllLastRun = vi.fn()
+const mockInvalidateRunHistory = vi.fn()
 vi.mock('@/service/use-workflow', () => ({
   useInvalidAllLastRun: () => mockInvalidAllLastRun,
+  useInvalidateWorkflowRunHistory: () => mockInvalidateRunHistory,
 }))
 
 // Mock FlowType
@@ -472,6 +474,7 @@ describe('usePipelineRun', () => {
       })
 
       expect(onWorkflowStarted).toHaveBeenCalledWith({ task_id: 'task-1' })
+      expect(mockInvalidateRunHistory).toHaveBeenCalled()
     })
 
     it('should call onWorkflowFinished callback when provided', async () => {
@@ -493,6 +496,7 @@ describe('usePipelineRun', () => {
       })
 
       expect(onWorkflowFinished).toHaveBeenCalledWith({ status: 'succeeded' })
+      expect(mockInvalidateRunHistory).toHaveBeenCalled()
     })
 
     it('should call onError callback when provided', async () => {
@@ -514,6 +518,7 @@ describe('usePipelineRun', () => {
       })
 
       expect(onError).toHaveBeenCalledWith({ message: 'error' })
+      expect(mockInvalidateRunHistory).toHaveBeenCalled()
     })
 
     it('should call onNodeStarted callback when provided', async () => {

web/eslint-suppressions.json

@@ -2299,7 +2299,7 @@
   },
   "app/components/base/markdown-blocks/code-block.tsx": {
     "react-hooks-extra/no-direct-set-state-in-use-effect": {
-      "count": 10
+      "count": 7
     },
     "tailwindcss/enforce-consistent-class-order": {
       "count": 1

web/package.json

@@ -1,7 +1,7 @@
 {
   "name": "dify-web",
   "type": "module",
-  "version": "1.12.1",
+  "version": "1.13.0",
   "private": true,
   "packageManager": "pnpm@10.27.0+sha512.72d699da16b1179c14ba9e64dc71c9a40988cbdc65c264cb0e489db7de917f20dcf4d64d8723625f2969ba52d4b7e2a1170682d9ac2a5dcaeaab732b7e16f04a",
   "imports": {