diff --git a/api/app_factory.py b/api/app_factory.py
index 52ae05583a..4fbd182e11 100644
--- a/api/app_factory.py
+++ b/api/app_factory.py
@@ -30,6 +30,8 @@ def create_app() -> DifyApp:
     start_time = time.perf_counter()
     app = create_flask_app_with_configs()
     initialize_extensions(app)
+    from initializer import run_initializers
+    run_initializers(app)
     end_time = time.perf_counter()
     if dify_config.DEBUG:
         logging.info(f"Finished create_app ({round((end_time - start_time) * 1000, 2)} ms)")
diff --git a/api/configs/apo/__init__.py b/api/configs/apo/__init__.py
index 7d21a27594..3398840850 100644
--- a/api/configs/apo/__init__.py
+++ b/api/configs/apo/__init__.py
@@ -14,4 +14,12 @@ class APOConfig(BaseSettings):
     APO_VM_URL: str = Field(
         description="apo vm url",
         default="http://localhost:8080",
+    )
+    INITIAL_LANGUAGE: str = Field(
+        description="Initial workflows' language",
+        default="en-US"
+    )
+    WORKFLOW_DIR: str = Field(
+        description="Directory of workflows yaml file.",
+        default="./workflows"
     )
\ No newline at end of file
diff --git a/api/core/tools/builtin_tool/providers/apo_select/tools/container_cpu_throttle_containerd_seconds.py b/api/core/tools/builtin_tool/providers/apo_select/tools/container_cpu_throttle_containerd_seconds.py
new file mode 100644
index 0000000000..78d9485d80
--- /dev/null
+++ b/api/core/tools/builtin_tool/providers/apo_select/tools/container_cpu_throttle_containerd_seconds.py
@@ -0,0 +1,48 @@
+import json
+from collections.abc import Generator
+from typing import Any, Optional
+
+import requests
+
+from configs import dify_config
+from core.tools.builtin_tool.tool import BuiltinTool
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from libs.apo_utils import APOUtils
+
+
+class ContainerCpuThrottleContainerdSecondsTool(BuiltinTool):
+    def _invoke(
+        self,
+        user_id: str,
+        tool_parameters: dict[str, Any],
+        conversation_id: Optional[str] = None,
+        app_id: Optional[str] = None,
+        message_id: Optional[str] = None,
+    ) -> Generator[ToolInvokeMessage, None, None]:
+        cadvisor_job_name = tool_parameters.get('cadvisor_job_name', '.*')
+        namespace = tool_parameters.get('namespace', '.*')
+        pod = tool_parameters.get('pod', '.*')
+        start_time = tool_parameters.get("startTime")
+        end_time = tool_parameters.get("endTime")
+        params = {
+            'metricName': '基础设施情况 - 容器CPU - 容器CPU节流时间 - Containerd',
+            'params': {
+                'cadvisor_job_name': cadvisor_job_name,
+                'namespace': namespace,
+                'pod': pod
+            },
+            'startTime': start_time,
+            'endTime': end_time,
+            'step': APOUtils.get_step(start_time, end_time),
+        }
+        resp = requests.post(dify_config.APO_BACKEND_URL + '/api/metric/query', json=params)
+        list = resp.json()['result']
+        list = json.dumps({
+            'type': 'metric',
+            'display': True,
+            'unit': list['unit'],
+            'data': {
+                'timeseries': list['timeseries']
+            }
+        })
+        yield self.create_text_message(list)
\ No newline at end of file
diff --git a/api/core/tools/builtin_tool/providers/apo_select/tools/container_cpu_throttle_containerd_seconds.yaml b/api/core/tools/builtin_tool/providers/apo_select/tools/container_cpu_throttle_containerd_seconds.yaml
new file mode 100644
index 0000000000..7835a3292e
--- /dev/null
+++ b/api/core/tools/builtin_tool/providers/apo_select/tools/container_cpu_throttle_containerd_seconds.yaml
@@ -0,0 +1,75 @@
+identity:
+  name: 容器CPU节流时长(使用Containerd容器运行时,按容器和Pod统计)
+  author: APO
+  label:
+    en_US: Container CPU throttling time (Containerd runtime, aggregated by container and Pod)
+    zh_Hans: 容器CPU节流时长(使用Containerd容器运行时,按容器和Pod统计)
+description:
+  human:
+    en_US: Container CPU throttling time (Containerd runtime, aggregated by container and Pod)
+    zh_Hans: 容器CPU节流时长(使用Containerd容器运行时,按容器和Pod统计)
+  llm: Container CPU throttling time (Containerd runtime, aggregated by container and Pod)
+display:
+  type: metric
+  title: 基础设施情况 - 容器CPU - 容器CPU节流时间 - Containerd
+  unit: "s"
+parameters:
+  - name: cadvisor_job_name
+    type: string
+    required: False
+    label:
+      en_US: cAdvisor job name
+      zh_Hans: cAdvisor任务名称
+    human_description:
+      en_US: cAdvisor job name
+      zh_Hans: cAdvisor任务名称
+    llm_description: cAdvisor job name
+    form: llm
+  - name: namespace
+    type: string
+    required: False
+    label:
+      en_US: Namespace
+      zh_Hans: 命名空间
+    human_description:
+      en_US: Namespace
+      zh_Hans: 命名空间
+    llm_description: Namespace
+    form: llm
+  - name: pod
+    type: string
+    required: False
+    label:
+      en_US: Pod name
+      zh_Hans: Pod名称
+    human_description:
+      en_US: Pod name
+      zh_Hans: Pod名称
+    llm_description: Pod name
+    form: llm
+  - name: startTime
+    type: number
+    required: true
+    label:
+      en_US: startTime
+      zh_Hans: startTime
+      pt_BR: startTime
+    human_description:
+      en_US: Data query start time
+      zh_Hans: 开始时间 (微秒)
+      pt_BR: Data query start time
+    llm_description: Data query start time
+    form: llm
+  - name: endTime
+    type: number
+    required: true
+    label:
+      en_US: endTime
+      zh_Hans: endTime
+      pt_BR: endTime
+    human_description:
+      en_US: Data query end time
+      zh_Hans: 结束时间 (微秒)
+      pt_BR: Data query end time
+    llm_description: Data query start time
+    form: llm
diff --git a/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_receive.py b/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_receive.py
new file mode 100644
index 0000000000..71509dd132
--- /dev/null
+++ b/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_receive.py
@@ -0,0 +1,44 @@
+import json
+from collections.abc import Generator
+from typing import Any, Optional
+
+import requests
+
+from configs import dify_config
+from core.tools.builtin_tool.tool import BuiltinTool
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from libs.apo_utils import APOUtils
+
+
+class LinuxNetworkDroppedPacketsReceiveTool(BuiltinTool):
+    def _invoke(
+        self,
+        user_id: str,
+        tool_parameters: dict[str, Any],
+        conversation_id: Optional[str] = None,
+        app_id: Optional[str] = None,
+        message_id: Optional[str] = None,
+    ) -> Generator[ToolInvokeMessage, None, None]:
+        cluster = tool_parameters.get('cluster', '.*')
+        start_time = tool_parameters.get("startTime")
+        end_time = tool_parameters.get("endTime")
+        params = {
+            'metricName': '集群总览 - 总览 - 网络饱和 - 丢包数 - Linux Packets dropped (receive)',
+            'params': {
+                'cluster': cluster
+            },
+            'startTime': start_time,
+            'endTime': end_time,
+            'step': APOUtils.get_step(start_time, end_time),
+        }
+        resp = requests.post(dify_config.APO_BACKEND_URL + '/api/metric/query', json=params)
+        list = resp.json()['result']
+        list = json.dumps({
+            'type': 'metric',
+            'display': True,
+            'unit': list['unit'],
+            'data': {
+                'timeseries': list['timeseries']
+            }
+        })
+        yield self.create_text_message(list)
\ No newline at end of file
diff --git a/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_receive.yaml b/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_receive.yaml
new file mode 100644
index 0000000000..0091db1961
--- /dev/null
+++ b/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_receive.yaml
@@ -0,0 +1,53 @@
+identity:
+  name: Linux集群网络接收丢包数
+  author: APO
+  label:
+    en_US: Linux cluster network receive packet drops
+    zh_Hans: Linux集群网络接收丢包数
+description:
+  human:
+    en_US: Linux cluster network receive packet drops
+    zh_Hans: Linux集群网络接收丢包数
+  llm: Linux cluster network receive packet drops
+display:
+  type: metric
+  title: 集群总览 - 总览 - 网络饱和 - 丢包数 - Linux Packets dropped (receive)
+  unit: "short"
+parameters:
+  - name: cluster
+    type: string
+    required: False
+    label:
+      en_US: Cluster name
+      zh_Hans: 集群名称
+    human_description:
+      en_US: Cluster name
+      zh_Hans: 集群名称
+    llm_description: Cluster name
+    form: llm
+  - name: startTime
+    type: number
+    required: true
+    label:
+      en_US: startTime
+      zh_Hans: startTime
+      pt_BR: startTime
+    human_description:
+      en_US: Data query start time
+      zh_Hans: 开始时间 (微秒)
+      pt_BR: Data query start time
+    llm_description: Data query start time
+    form: llm
+  - name: endTime
+    type: number
+    required: true
+    label:
+      en_US: endTime
+      zh_Hans: endTime
+      pt_BR: endTime
+    human_description:
+      en_US: Data query end time
+      zh_Hans: 结束时间 (微秒)
+      pt_BR: Data query end time
+    llm_description: Data query start time
+    form: llm
diff --git a/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_transmit.py b/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_transmit.py
new file mode 100644
index 0000000000..c62f723e7f
--- /dev/null
+++ b/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_transmit.py
@@ -0,0 +1,44 @@
+import json
+from collections.abc import Generator
+from typing import Any, Optional
+
+import requests
+
+from configs import dify_config
+from core.tools.builtin_tool.tool import BuiltinTool
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from libs.apo_utils import APOUtils
+
+
+class LinuxNetworkDroppedPacketsTransmitTool(BuiltinTool):
+    def _invoke(
+        self,
+        user_id: str,
+        tool_parameters: dict[str, Any],
+        conversation_id: Optional[str] = None,
+        app_id: Optional[str] = None,
+        message_id: Optional[str] = None,
+    ) -> Generator[ToolInvokeMessage, None, None]:
+        cluster = tool_parameters.get('cluster', '.*')
+        start_time = tool_parameters.get("startTime")
+        end_time = tool_parameters.get("endTime")
+        params = {
+            'metricName': '集群总览 - 总览 - 网络饱和 - 丢包数 - Linux Packets dropped (transmit)',
+            'params': {
+                'cluster': cluster
+            },
+            'startTime': start_time,
+            'endTime': end_time,
+            'step': APOUtils.get_step(start_time, end_time),
+        }
+        resp = requests.post(dify_config.APO_BACKEND_URL + '/api/metric/query', json=params)
+        list = resp.json()['result']
+        list = json.dumps({
+            'type': 'metric',
+            'display': True,
+            'unit': list['unit'],
+            'data': {
+                'timeseries': list['timeseries']
+            }
+        })
+        yield self.create_text_message(list)
\ No newline at end of file
diff --git a/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_transmit.yaml b/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_transmit.yaml
new file mode 100644
index 0000000000..bf7c9bcff8
--- /dev/null
+++ b/api/core/tools/builtin_tool/providers/apo_select/tools/linux_network_dropped_packets_transmit.yaml
@@ -0,0 +1,53 @@
+identity:
+  name: Linux集群网络发送丢包数
+  author: APO
+  label:
+    en_US: Linux cluster network transmit packet drops
+    zh_Hans: Linux集群网络发送丢包数
+description:
+  human:
+    en_US: Linux cluster network transmit packet drops
+    zh_Hans: Linux集群网络发送丢包数
+  llm: Linux cluster network transmit packet drops
+display:
+  type: metric
+  title: 集群总览 - 总览 - 网络饱和 - 丢包数 - Linux Packets dropped (transmit)
+  unit: "short"
+parameters:
+  - name: cluster
+    type: string
+    required: False
+    label:
+      en_US: Cluster name
+      zh_Hans: 集群名称
+    human_description:
+      en_US: Cluster name
+      zh_Hans: 集群名称
+    llm_description: Cluster name
+    form: llm
+  - name: startTime
+    type: number
+    required: true
+    label:
+      en_US: startTime
+      zh_Hans: startTime
+      pt_BR: startTime
+    human_description:
+      en_US: Data query start time
+      zh_Hans: 开始时间 (微秒)
+      pt_BR: Data query start time
+    llm_description: Data query start time
+    form: llm
+  - name: endTime
+    type: number
+    required: true
+    label:
+      en_US: endTime
+      zh_Hans: endTime
+      pt_BR: endTime
+    human_description:
+      en_US: Data query end time
+      zh_Hans: 结束时间 (微秒)
+      pt_BR: Data query end time
+    llm_description: Data query start time
+    form: llm
diff --git a/api/core/tools/builtin_tool/providers/apo_select/tools/node_cpu_utilization_linux.py b/api/core/tools/builtin_tool/providers/apo_select/tools/node_cpu_utilization_linux.py
new file mode 100644
index 0000000000..618c0bb3be
--- /dev/null
+++ b/api/core/tools/builtin_tool/providers/apo_select/tools/node_cpu_utilization_linux.py
@@ -0,0 +1,44 @@
+import json
+from collections.abc import Generator
+from typing import Any, Optional
+
+import requests
+
+from configs import dify_config
+from core.tools.builtin_tool.tool import BuiltinTool
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from libs.apo_utils import APOUtils
+
+
+class NodeCpuUtilizationLinuxTool(BuiltinTool):
+    def _invoke(
+        self,
+        user_id: str,
+        tool_parameters: dict[str, Any],
+        conversation_id: Optional[str] = None,
+        app_id: Optional[str] = None,
+        message_id: Optional[str] = None,
+    ) -> Generator[ToolInvokeMessage, None, None]:
+        cluster = tool_parameters.get('cluster', '.*')
+        start_time = tool_parameters.get("startTime")
+        end_time = tool_parameters.get("endTime")
+        params = {
+            'metricName': '集群总览 - 节点资源使用 - 节点CPU使用率 - Linux',
+            'params': {
+                'cluster': cluster
+            },
+            'startTime': start_time,
+            'endTime': end_time,
+            'step': APOUtils.get_step(start_time, end_time),
+        }
+        resp = requests.post(dify_config.APO_BACKEND_URL + '/api/metric/query', json=params)
+        list = resp.json()['result']
+        list = json.dumps({
+            'type': 'metric',
+            'display': True,
+            'unit': list['unit'],
+            'data': {
+                'timeseries': list['timeseries']
+            }
+        })
+        yield self.create_text_message(list)
\ No newline at end of file
diff --git a/api/core/tools/builtin_tool/providers/apo_select/tools/node_cpu_utilization_linux.yaml b/api/core/tools/builtin_tool/providers/apo_select/tools/node_cpu_utilization_linux.yaml
new file mode 100644
index 0000000000..d5014fa446
--- /dev/null
+++ b/api/core/tools/builtin_tool/providers/apo_select/tools/node_cpu_utilization_linux.yaml
@@ -0,0 +1,53 @@
+identity:
+  name: Linux节点CPU使用率
+  author: APO
+  label:
+    en_US: Linux node CPU utilization rate
+    zh_Hans: Linux节点CPU使用率
+description:
+  human:
+    en_US: Linux node CPU utilization rate
+    zh_Hans: Linux节点CPU使用率
+  llm: Linux node CPU utilization rate
+display:
+  type: metric
+  title: 集群总览 - 节点资源使用 - 节点CPU使用率 - Linux
+  unit: "percentunit"
+parameters:
+  - name: cluster
+    type: string
+    required: False
+    label:
+      en_US: Cluster name
+      zh_Hans: 集群名称
+    human_description:
+      en_US: Cluster name
+      zh_Hans: 集群名称
+    llm_description: Cluster name
+    form: llm
+  - name: startTime
+    type: number
+    required: true
+    label:
+      en_US: startTime
+      zh_Hans: startTime
+      pt_BR: startTime
+    human_description:
+      en_US: Data query start time
+      zh_Hans: 开始时间 (微秒)
+      pt_BR: Data query start time
+    llm_description: Data query start time
+    form: llm
+  - name: endTime
+    type: number
+    required: true
+    label:
+      en_US: endTime
+      zh_Hans: endTime
+      pt_BR: endTime
+    human_description:
+      en_US: Data query end time
+      zh_Hans: 结束时间 (微秒)
+      pt_BR: Data query end time
+    llm_description: Data query start time
+    form: llm
diff --git a/api/core/tools/builtin_tool/providers/apo_select/tools/node_memory_usage_linux.py b/api/core/tools/builtin_tool/providers/apo_select/tools/node_memory_usage_linux.py
new file mode 100644
index 0000000000..8b0608f33b
--- /dev/null
+++ b/api/core/tools/builtin_tool/providers/apo_select/tools/node_memory_usage_linux.py
@@ -0,0 +1,44 @@
+import json
+from collections.abc import Generator
+from typing import Any, Optional
+
+import requests
+
+from configs import dify_config
+from core.tools.builtin_tool.tool import BuiltinTool
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from libs.apo_utils import APOUtils
+
+
+class NodeMemoryUsageLinuxTool(BuiltinTool):
+    def _invoke(
+        self,
+        user_id: str,
+        tool_parameters: dict[str, Any],
+        conversation_id: Optional[str] = None,
+        app_id: Optional[str] = None,
+        message_id: Optional[str] = None,
+    ) -> Generator[ToolInvokeMessage, None, None]:
+        cluster = tool_parameters.get('cluster', '.*')
+        start_time = tool_parameters.get("startTime")
+        end_time = tool_parameters.get("endTime")
+        params = {
+            'metricName': '集群总览 - 节点资源使用 - 节点内存使用量 - Linux',
+            'params': {
+                'cluster': cluster
+            },
+            'startTime': start_time,
+            'endTime': end_time,
+            'step': APOUtils.get_step(start_time, end_time),
+        }
+        resp = requests.post(dify_config.APO_BACKEND_URL + '/api/metric/query', json=params)
+        list = resp.json()['result']
+        list = json.dumps({
+            'type': 'metric',
+            'display': True,
+            'unit': list['unit'],
+            'data': {
+                'timeseries': list['timeseries']
+            }
+        })
+        yield self.create_text_message(list)
\ No newline at end of file
diff --git a/api/core/tools/builtin_tool/providers/apo_select/tools/node_memory_usage_linux.yaml b/api/core/tools/builtin_tool/providers/apo_select/tools/node_memory_usage_linux.yaml
new file mode 100644
index 0000000000..d41a224ceb
--- /dev/null
+++ b/api/core/tools/builtin_tool/providers/apo_select/tools/node_memory_usage_linux.yaml
@@ -0,0 +1,53 @@
+identity:
+  name: Linux节点内存使用字节数
+  author: APO
+  label:
+    en_US: Linux node memory usage in bytes
+    zh_Hans: Linux节点内存使用字节数
+description:
+  human:
+    en_US: Linux node memory usage in bytes
+    zh_Hans: Linux节点内存使用字节数
+  llm: Linux node memory usage in bytes
+display:
+  type: metric
+  title: 集群总览 - 节点资源使用 - 节点内存使用量 - Linux
+  unit: "bytes"
+parameters:
+  - name: cluster
+    type: string
+    required: False
+    label:
+      en_US: Cluster name
+      zh_Hans: 集群名称
+    human_description:
+      en_US: Cluster name
+      zh_Hans: 集群名称
+    llm_description: Cluster name
+    form: llm
+  - name: startTime
+    type: number
+    required: true
+    label:
+      en_US: startTime
+      zh_Hans: startTime
+      pt_BR: startTime
+    human_description:
+      en_US: Data query start time
+      zh_Hans: 开始时间 (微秒)
+      pt_BR: Data query start time
+    llm_description: Data query start time
+    form: llm
+  - name: endTime
+    type: number
+    required: true
+    label:
+      en_US: endTime
+      zh_Hans: endTime
+      pt_BR: endTime
+    human_description:
+      en_US: Data query end time
+      zh_Hans: 结束时间 (微秒)
+      pt_BR: Data query end time
+    llm_description: Data query start time
+    form: llm
diff --git a/api/initializer/__init__.py b/api/initializer/__init__.py
new file mode 100644
index 0000000000..8a6c278d5a
--- /dev/null
+++ b/api/initializer/__init__.py
@@ -0,0 +1,11 @@
+from flask import Flask
+from . import (
+    account,
+    workflow
+)
+from .decorator import _initializers
+
+def run_initializers(app: Flask):
+    with app.app_context():
+        for func, _ in sorted(_initializers, key=lambda x: x[1]):
+            func()
\ No newline at end of file
diff --git a/api/initializer/account.py b/api/initializer/account.py
new file mode 100644
index 0000000000..c6bf1e43c0
--- /dev/null
+++ b/api/initializer/account.py
@@ -0,0 +1,28 @@
+from models import Account
+from extensions.ext_database import db
+from services.account_service import RegisterService, AccountService, TenantService
+from .decorator import initializer
+from .admin import get_admin
+
+
+@initializer(priority=1)
+def init_admin_account():
+    if db.session.query(Account).filter_by(name="admin").first():
+        return
+    
+    registerService = RegisterService()
+    registerService.setup("admin@apo.com", "admin", "APO2024@admin", "")
+
+
+@initializer(priority=2)
+def init_anonymous_account():
+    if db.session.query(Account).filter_by(name="anonymous").first():
+        return
+    
+    accountService = AccountService()
+    anonymous = accountService.create_account("anonymous@apo.com", "anonymous", "en-US", "APO2024@anonymous", True)
+
+    admin = get_admin()
+    tenantService = TenantService()
+    tenant = tenantService.get_current_tenant_by_account(admin)
+    tenantService.create_tenant_member(tenant, anonymous)
\ No newline at end of file
diff --git a/api/initializer/admin.py b/api/initializer/admin.py
new file mode 100644
index 0000000000..66ab05b1e6
--- /dev/null
+++ b/api/initializer/admin.py
@@ -0,0 +1,11 @@
+from extensions.ext_database import db
+from models import Account, Tenant, TenantAccountJoin, TenantAccountRole
+
+def get_admin() -> Account:
+    admin = db.session.query(Account).filter_by(name="admin").first()
+    tenant_account_join = db.session.query(TenantAccountJoin).filter_by(account_id=admin.id, role=TenantAccountRole.OWNER).first()
+    tenant = db.session.query(Tenant).filter_by(id=tenant_account_join.tenant_id).first()
+    admin.current_tenant = tenant
+    admin.current_tenant_id = tenant.id
+    return admin
+    
\ No newline at end of file
diff --git a/api/initializer/decorator.py b/api/initializer/decorator.py
new file mode 100644
index 0000000000..38d954aada
--- /dev/null
+++ b/api/initializer/decorator.py
@@ -0,0 +1,9 @@
+from typing import Callable, List, Tuple
+
+_initializers: List[Tuple[Callable, int]] = []
+
+def initializer(priority: int = 10) -> Callable:
+    def decorator(func: Callable) -> Callable:
+        _initializers.append((func, priority))
+        return func
+    return decorator
diff --git a/api/initializer/workflow.py b/api/initializer/workflow.py
new file mode 100644
index 0000000000..2ca0dc0b01
--- /dev/null
+++ b/api/initializer/workflow.py
@@ -0,0 +1,203 @@
+import os
+import yaml
+import logging
+
+from sqlalchemy.exc import SQLAlchemyError
+from services.app_dsl_service import AppDslService
+from services.workflow_service import WorkflowService
+from extensions.ext_database import db
+from .decorator import initializer
+from .admin import get_admin
+from models import App, ApiToken, Workflow, InstalledApp
+from contexts import tenant_id
+from configs.app_config import APOConfig
+from typing import Union
+
+@initializer(priority=3)
+def init_workflow():
+    apo_config = APOConfig()
+    initial_language = apo_config.INITIAL_LANGUAGE
+    workflow_dir = apo_config.WORKFLOW_DIR
+
+    sub_dir = ''
+    if initial_language == 'en-US':
+        sub_dir = 'en'
+    elif initial_language == 'zh-Hans':
+        sub_dir = 'zh'
+    workflows = []
+    dir = f'{workflow_dir}/{sub_dir}'
+    if not os.path.isdir(dir):
+        raise ValueError(f"Invalid directory: {dir}")
+    
+    for file_entry in os.scandir(dir):
+        if not file_entry.name.endswith('.yaml') and not file_entry.name.endswith('.yml') or file_entry.name.startswith('.'):
+            continue
+        try:
+            with open(file_entry.path, 'r', encoding='utf-8') as file:
+                content = file.read()
+                workflows.append(content)
+        except Exception as e:
+            logging.ERROR(f"Failed to read file: {file_entry.path}")
+    
+    admin = get_admin()
+    try:
+        original_token = tenant_id.set(admin.current_tenant_id) 
+        for w in workflows:
+            result = _check_workflow_to_update(db.session, w, admin)
+ 
+            if result is False:
+                continue
+
+            import_service = AppDslService(db.session)
+            workflow_service = WorkflowService()
+            
+            '''Import an app or update existing app'''
+            imp = import_service.import_app(
+                account=admin,
+                import_mode="yaml-content",
+                yaml_content=w,
+                app_id=result if result else None
+            )
+
+            app_model = (
+                db.session.query(App)
+                .filter(
+                    App.id == imp.app_id,
+                    App.tenant_id == admin.current_tenant_id,
+                    App.status == "normal"
+                )
+                .first()
+            )
+
+            workflow_service.publish_workflow(app_model=app_model, account=admin)
+            if result is None:
+                _generate_api_key(db.session, imp.app_id, admin)
+        db.session.commit()
+        _adjust_workflows(initial_language)
+    except Exception as e:
+        db.session.rollback()
+        raise
+    finally:
+        tenant_id.reset(original_token)
+
+def _check_workflow_to_update(session, content, account) -> Union[str, bool, None]:
+    """Check if the workflow needs to be updated or created.
+    - None: Need to create a new App.
+    - str(app_id): Need to update the existing App.
+    - False: No need to update.
+    """
+    try:
+        content_dict = yaml.safe_load(content)
+        app_name = content_dict.get('app', {}).get('name')
+        graph = content_dict.get('workflow', {}).get('graph')
+        features = content_dict.get('workflow', {}).get('features')
+        
+        app_model = (
+            session.query(App)
+            .filter(
+                App.name == app_name,
+                App.tenant_id == account.current_tenant_id,
+                App.status == "normal",
+            )
+            .first()
+        )
+
+        if not app_model:
+            return None
+
+        workflow = (
+            session.query(Workflow)
+            .filter(
+                Workflow.graph == graph,
+                Workflow.features == features,
+            )
+            .first()
+        )
+
+        if workflow:
+            return False
+        else:
+            return app_model.id
+
+    except Exception as e:
+        logging.error(f"Failed to check workflow: {str(e)}")
+        return False
+
+def _generate_api_key(session, app_id, account, key=None):
+    if not app_id or not account:
+        return
+    
+    key = key or ApiToken.generate_api_key('app-', 24)
+    api_token = ApiToken(
+        app_id=app_id,
+        tenant_id=account.current_tenant_id,
+        token=key,
+        type='app'
+    )
+    session.add(api_token)
+
+
+def _adjust_workflows(language):
+    to_adjust = {
+        "zh-Hans": {
+            "告警有效性分析": {
+                "app_id": "dcfeddd2-d6e7-4dc4-a284-e48ab56bf6af",
+                "api_token": "app-x0mOJKUvhr35BOISSeNmsfXj"
+            },
+            "告警简单根因分析": {
+                "app_id": "a2d4d3aa-3401-4393-859e-df051bdd5cd1"
+            }
+        },
+        "en-US": {
+            "alert validity confirmation": {
+                "app_id": "dcfeddd2-d6e7-4dc4-a284-e48ab56bf6af",
+                "api_token": "app-x0mOJKUvhr35BOISSeNmsfXj"
+            },
+            "alert simple root cause analysis": {
+                "app_id": "a2d4d3aa-3401-4393-859e-df051bdd5cd1"
+            }
+        }
+    }
+    try:
+        config = to_adjust.get(language)
+        for app_name, field_config in config.items():
+                app = db.session.query(App).filter_by(name=app_name).first()
+                new_app_id = field_config["app_id"]
+                if not app:
+                    continue
+                
+                if app.id == new_app_id:
+                    continue
+
+                origin_app_id = app.id
+                update_mappings = [
+                    (ApiToken, {"app_id": new_app_id}),
+                    (Workflow, {"app_id": new_app_id}),
+                    (InstalledApp, {"app_id": new_app_id})
+                ]
+
+                for model, update_values in update_mappings:
+                    db.session.query(model)\
+                        .filter_by(app_id=origin_app_id)\
+                        .update(update_values, synchronize_session=False)
+
+                app.id = new_app_id
+                db.session.merge(app)
+
+                if "api_token" in field_config:
+                    new_token = field_config["api_token"]
+                    db.session.query(ApiToken)\
+                        .filter_by(app_id=new_app_id)\
+                        .update({"token": new_token}, synchronize_session=False)
+
+                db.session.commit()
+                print(f"Successfully updated {app_name} (ID: {origin_app_id} → {new_app_id})")
+
+    except SQLAlchemyError as e:
+        db.session.rollback()
+        logging.ERROR(f"Database error occurred: {str(e)}")
+        raise
+    except Exception as e:
+        db.session.rollback()
+        logging.ERROR(f"Unexpected error: {str(e)}")
+        raise
\ No newline at end of file