feat(tenancy): add Workspace multi-tenant foundation (#2353)

* Document multi-tenant workspace architecture * Add OSS and commercial workspace boundaries * docs: redesign multi-tenant workspace architecture * feat(tenancy): implement workspace isolation * docs(tenancy): record verification evidence * docs(tenancy): revise single-instance SaaS topology * docs(tenancy): refine architecture options * docs: finalize cloud v2 multi-tenant decisions * feat(tenancy): establish cloud isolation foundations * feat(tenancy): harden shared cloud runtime boundaries * docs(tenancy): record final isolation verification * fix(tenancy): close isolation and permission gaps * docs(tenancy): record final isolation verification * feat(tenancy): connect cloud workspace control plane * fix(build): install git for pinned SDK * docs(cloud): update control plane verification * chore: update multi-tenant SDK pin * fix(cloud): skip legacy model sync during startup * test(cloud): preserve minimal model manager fixtures * fix(cloud): preserve authenticated account context * fix(cloud): reuse authenticated account for user info * feat(cloud): complete Workspace settings navigation * test(web): cover Workspace dropdown menu * feat(web): place workspace controls in sidebar * refactor(web): streamline workspace controls * style(web): format workspace layout test * fix(cloud): surface runtime and workspace plan status * fix(plugin): keep runtime identity stable across restarts * fix(ui): widen and center workspace switcher * fix(ui): hide roles from workspace switcher * fix(ui): align workspace switcher with sidebar entries * feat(workspace): add in-product collaboration and direct Cloud launch * style: format collaboration changes * fix(workspace): bind collaboration APIs to tenant UoW * fix(cloud): preserve Core-owned collaboration state * test(cloud): require Space identity for invite registration * feat(cloud): complete secure invitation experience * style(web): format invitation flows * fix(cloud): recover box runtime without unscoped skill reload * feat(oss): enforce invitation account and owner billing flows * style: format OSS account service * test(oss): cover invitation logout handoff * fix(oss): resolve workspace owner in scoped session * feat(cloud): harden multi-tenant runtime resources * fix(cloud): bound runtime restart storms * fix(cloud): eliminate periodic runtime CPU spikes * fix(cloud): enforce instance capacity ceilings * fix(cloud): scope public login capability discovery * fix(cloud): bound tenant maintenance and monitoring work * fix(runtime): bound tenant resource amplification * fix(deps): pin green multi-tenant plugin SDK * fix(cloud): handle unavailable skill capability * fix(security): require authentication for image file endpoint (H-2) - Changed /api/v1/files/image from AuthType.NONE to USER_TOKEN_OR_API_KEY - Added Permission.RESOURCE_VIEW requirement - Prevents unauthenticated cross-tenant file access via leaked keys - Fixes HIGH severity finding from multi-tenant security review docs: add comprehensive database migration guide - Complete migration steps for OSS → multi-tenant - Backup, execution, verification procedures - Rollback scenarios and recovery plans - Performance tuning recommendations * test: add comprehensive cross-tenant isolation tests Added 7 critical test scenarios for multi-tenant boundaries: - Cross-tenant bot access prevention - Viewer role read-only enforcement - Removed member immediate access revocation - Model provider credential isolation - WebSocket message isolation - Invitation token workspace scoping - Multi-workspace context validation These tests address P0-2 coverage gaps for: - workspaces.py (membership & invitation flows) - user.py (authentication & authorization) - websocket_chat.py (real-time isolation) - plugins.py (resource access control) docs: finalize database migration guide * fix(security): resolve M-1, M-2, M-3 security findings M-1: WebSocket authorization TOCTOU race (FIXED) - Changed _revalidate_websocket_authorization to return RequestContext - Ensures validated context is used immediately without race window - Prevents removed members from sending messages during revalidation gap M-2: Model Manager cache workspace isolation (VERIFIED) - Confirmed _CacheKey already uses 4-tuple: (instance, workspace, generation, resource) - Cache is properly scoped per workspace, no cross-tenant leakage possible - No code change needed, documented as working correctly M-3: Invitation lock workspace scoping (FIXED) - Changed lock key from token_digest to workspace_uuid:token_digest - Prevents DoS where attacker locks token in Workspace A to block Workspace B - Locks now isolated per workspace All MEDIUM severity findings from security review now resolved. * fix(cloud): unblock tenant CI and enforce knowledge quotas * fix(tenancy): scope rerank model sync --------- Co-authored-by: dadachann <185672915+dadachann@users.noreply.github.com>
2026-08-09 12:40:59 +00:00 · 2026-07-30 21:43:35 +08:00
parent 463b120923
commit e1ac5e0fc8
468 changed files with 78320 additions and 13137 deletions
@@ -6,6 +6,7 @@ Tests the helper methods that don't require real Dify API calls.
 from __future__ import annotations

 import pytest
+import time
 from unittest.mock import AsyncMock, MagicMock

 import langbot_plugin.api.entities.builtin.platform.message as platform_message
@@ -18,13 +19,12 @@ class TestDifyWorkflowSubmitClient:

        class FakeResponse:
            status_code = 503
+            headers = {}

-            async def aread(self):
-                return b''
-
-            async def aiter_lines(self):
-                raise AssertionError('error responses must not enter the SSE loop')
-                yield
+            async def aiter_bytes(self, chunk_size=None):
+                del chunk_size
+                if False:
+                    yield b''

        class FakeStreamContext:
            async def __aenter__(self):
@@ -66,6 +66,33 @@ class TestDifyWorkflowSubmitClient:
                )
            )

+    @pytest.mark.asyncio
+    async def test_sse_parser_rejects_an_unbounded_line(self):
+        from langbot.libs.dify_service_api.v1 import client, errors
+
+        class FakeResponse:
+            async def aiter_bytes(self, chunk_size=None):
+                del chunk_size
+                for _ in range(129):
+                    yield b'x' * 8192
+
+        with pytest.raises(errors.DifyAPIError, match='SSE event exceeds'):
+            await anext(client._iter_sse_json(FakeResponse()))
+
+    @pytest.mark.asyncio
+    async def test_upload_rejects_oversized_local_file(self, tmp_path):
+        from langbot.libs.dify_service_api.v1 import client
+
+        file_path = tmp_path / 'large.bin'
+        file_path.write_bytes(b'x' * (client._MAX_DIFY_UPLOAD_BYTES + 1))
+        dify_client = client.AsyncDifyServiceClient(
+            'test-key',
+            'https://dify.example/v1',
+        )
+
+        with pytest.raises(ValueError, match='exceeds the size limit'):
+            await dify_client.upload_file(file_path, 'person_user-1')
+

 class TestDifyExtractTextOutput:
    """Tests for _extract_dify_text_output method."""
@@ -252,42 +279,65 @@ class TestDifyHumanInputForms:
        runner.dify_client.upload_file = AsyncMock(return_value={'id': 'upload-1'})
        return runner

-    def test_pending_forms_are_isolated_by_bot_and_pipeline(self):
+    def test_pending_forms_are_isolated_by_workspace_generation_bot_and_pipeline(self):
        from langbot.pkg.provider.runners import difysvapi

        query_a = MagicMock()
+        query_a.instance_uuid = 'instance-a'
+        query_a.workspace_uuid = 'workspace-a'
+        query_a.placement_generation = 1
        query_a.bot_uuid = 'bot-a'
        query_a.pipeline_uuid = 'pipeline-a'
        query_a.session.launcher_type.value = 'person'
        query_a.session.launcher_id = 'shared-user'

        query_b = MagicMock()
+        query_b.instance_uuid = 'instance-a'
+        query_b.workspace_uuid = 'workspace-a'
+        query_b.placement_generation = 1
        query_b.bot_uuid = 'bot-b'
        query_b.pipeline_uuid = 'pipeline-a'
        query_b.session.launcher_type.value = 'person'
        query_b.session.launcher_id = 'shared-user'

        query_c = MagicMock()
+        query_c.instance_uuid = 'instance-a'
+        query_c.workspace_uuid = 'workspace-a'
+        query_c.placement_generation = 1
        query_c.bot_uuid = 'bot-a'
        query_c.pipeline_uuid = 'pipeline-b'
        query_c.session.launcher_type.value = 'person'
        query_c.session.launcher_id = 'shared-user'

+        query_d = MagicMock()
+        query_d.instance_uuid = 'instance-a'
+        query_d.workspace_uuid = 'workspace-b'
+        query_d.placement_generation = 2
+        query_d.bot_uuid = 'bot-a'
+        query_d.pipeline_uuid = 'pipeline-a'
+        query_d.session.launcher_type.value = 'person'
+        query_d.session.launcher_id = 'shared-user'
+
        key_a = difysvapi._session_key_from_query(query_a)
        key_b = difysvapi._session_key_from_query(query_b)
        key_c = difysvapi._session_key_from_query(query_c)
+        key_d = difysvapi._session_key_from_query(query_d)
        difysvapi._PENDING_FORMS.clear()
        difysvapi._set_pending_form(key_a, {'form_token': 'token-a', 'workflow_run_id': 'run-a'})
        difysvapi._set_pending_form(key_b, {'form_token': 'token-b', 'workflow_run_id': 'run-b'})
        difysvapi._set_pending_form(key_c, {'form_token': 'token-c', 'workflow_run_id': 'run-c'})
+        difysvapi._set_pending_form(key_d, {'form_token': 'token-d', 'workflow_run_id': 'run-d'})

        assert key_a != key_b
        assert key_a != key_c
+        assert key_a != key_d
        assert difysvapi._get_pending_form_by_token(key_a, 'token-a') is not None
        assert difysvapi._get_pending_form_by_token(key_a, 'token-b') is None
        assert difysvapi._get_pending_form_by_token(key_a, 'token-c') is None
+        assert difysvapi._get_pending_form_by_token(key_a, 'token-d') is None
        assert difysvapi._get_pending_form_by_token(key_b, 'token-b') is not None
        assert difysvapi._get_pending_form_by_token(key_c, 'token-c') is not None
+        assert difysvapi._get_pending_form_by_token(key_d, 'token-d') is not None
        assert difysvapi._get_latest_pending_form(key_a)['workflow_run_id'] == 'run-a'
        assert difysvapi._get_latest_pending_form(key_b)['workflow_run_id'] == 'run-b'
        assert difysvapi._get_latest_pending_form(key_c)['workflow_run_id'] == 'run-c'
@@ -297,6 +347,130 @@ class TestDifyHumanInputForms:
        assert difysvapi._dify_user_from_query(query_a) == difysvapi._dify_user_from_query(query_c)
        difysvapi._PENDING_FORMS.clear()

+    def test_pending_form_lookup_does_not_scan_unrelated_sessions(self, monkeypatch):
+        from langbot.pkg.provider.runners import difysvapi
+
+        def session_key(index: int):
+            return (
+                'instance',
+                f'workspace-{index}',
+                1,
+                'bot',
+                'pipeline',
+                'adapter',
+                'person',
+                f'user-{index}',
+            )
+
+        difysvapi._PENDING_FORMS.clear()
+        for index in range(512):
+            difysvapi._set_pending_form(
+                session_key(index),
+                {
+                    'form_token': f'token-{index}',
+                    'workflow_run_id': f'run-{index}',
+                },
+            )
+
+        class NoGlobalIterationDict(dict):
+            def __iter__(self):
+                raise AssertionError('pending form lookup scanned all sessions')
+
+            def keys(self):
+                raise AssertionError('pending form lookup scanned all sessions')
+
+            def items(self):
+                raise AssertionError('pending form lookup scanned all sessions')
+
+            def values(self):
+                raise AssertionError('pending form lookup scanned all sessions')
+
+        guarded_forms = NoGlobalIterationDict(difysvapi._PENDING_FORMS)
+        monkeypatch.setattr(difysvapi, '_PENDING_FORMS', guarded_forms)
+
+        assert difysvapi._get_pending_form_by_token(session_key(511), 'token-511')['workflow_run_id'] == 'run-511'
+        difysvapi._set_pending_form(
+            session_key(512),
+            {'form_token': 'token-512', 'workflow_run_id': 'run-512'},
+        )
+        assert len(guarded_forms) == 513
+
+    def test_pending_form_expiry_heap_ignores_stale_overwrite_and_stays_bounded(self):
+        from langbot.pkg.provider.runners import difysvapi
+
+        session_key = (
+            'instance',
+            'workspace',
+            1,
+            'bot',
+            'pipeline',
+            'adapter',
+            'person',
+            'user',
+        )
+        difysvapi._PENDING_FORMS.clear()
+        now = time.time()
+        difysvapi._set_pending_form(
+            session_key,
+            {
+                'form_token': 'token',
+                'workflow_run_id': 'stale',
+                'expiration_time': now + 1,
+            },
+        )
+        for revision in range(500):
+            difysvapi._set_pending_form(
+                session_key,
+                {
+                    'form_token': 'token',
+                    'workflow_run_id': f'current-{revision}',
+                    'expiration_time': now + 3600 + revision,
+                },
+            )
+
+        difysvapi._prune_pending_forms(now + 2)
+
+        assert difysvapi._get_pending_form_by_token(session_key, 'token')['workflow_run_id'] == 'current-499'
+        assert difysvapi._PENDING_FORM_ACTIVE_COUNT == 1
+        assert len(difysvapi._PENDING_FORM_EXPIRY_HEAP) <= max(
+            difysvapi._PENDING_FORM_HEAP_COMPACT_FLOOR,
+            difysvapi._PENDING_FORM_ACTIVE_COUNT * difysvapi._PENDING_FORM_HEAP_MAX_MULTIPLIER,
+        )
+
+    def test_pending_form_capacity_evicts_earliest_session_without_full_scan(
+        self,
+        monkeypatch,
+    ):
+        from langbot.pkg.provider.runners import difysvapi
+
+        def session_key(index: int):
+            return (
+                'instance',
+                f'workspace-{index}',
+                1,
+                'bot',
+                'pipeline',
+                'adapter',
+                'person',
+                f'user-{index}',
+            )
+
+        difysvapi._PENDING_FORMS.clear()
+        monkeypatch.setattr(difysvapi, '_PENDING_FORM_MAX_SESSIONS', 2)
+        now = time.time()
+        for index, expires_in in ((1, 300), (2, 100), (3, 200)):
+            difysvapi._set_pending_form(
+                session_key(index),
+                {
+                    'form_token': f'token-{index}',
+                    'expiration_time': now + expires_in,
+                },
+            )
+
+        assert session_key(1) in difysvapi._PENDING_FORMS
+        assert session_key(2) not in difysvapi._PENDING_FORMS
+        assert session_key(3) in difysvapi._PENDING_FORMS
+
    def test_interactive_form_data_preserves_pipeline_uuid(self):
        from langbot.pkg.provider.runners import difysvapi

@@ -0,0 +1,49 @@
+from __future__ import annotations
+
+import pytest
+
+from langbot.libs.deerflow_api.client import (
+    ERROR_BODY_MAX_BYTES,
+    _read_error_body,
+)
+from langbot.libs.deerflow_api.errors import DeerFlowAPIError
+from langbot.pkg.provider.runners.langflowapi import (
+    _MAX_LANGFLOW_LINE_CHARS,
+    _MAX_LANGFLOW_RESPONSE_BYTES,
+    _iter_limited_lines,
+    _read_limited_response,
+)
+
+
+class _ChunkedResponse:
+    def __init__(self, chunks: list[bytes]):
+        self._chunks = chunks
+
+    async def aiter_bytes(self, chunk_size=None):
+        del chunk_size
+        for chunk in self._chunks:
+            yield chunk
+
+
+@pytest.mark.asyncio
+async def test_langflow_rejects_oversized_stream_event():
+    response = _ChunkedResponse([b'x' * (_MAX_LANGFLOW_LINE_CHARS + 1)])
+
+    with pytest.raises(ValueError, match='event exceeds'):
+        await anext(_iter_limited_lines(response))
+
+
+@pytest.mark.asyncio
+async def test_langflow_rejects_oversized_blocking_response():
+    response = _ChunkedResponse([b'x' * (_MAX_LANGFLOW_RESPONSE_BYTES + 1)])
+
+    with pytest.raises(ValueError, match='response exceeds'):
+        await _read_limited_response(response)
+
+
+@pytest.mark.asyncio
+async def test_deerflow_rejects_oversized_error_body():
+    response = _ChunkedResponse([b'x' * (ERROR_BODY_MAX_BYTES + 1)])
+
+    with pytest.raises(DeerFlowAPIError, match='response exceeds'):
+        await _read_error_body(response)
@@ -0,0 +1,62 @@
+from __future__ import annotations
+
+import asyncio
+import threading
+from unittest.mock import AsyncMock
+
+import pytest
+
+from langbot.pkg.provider import runner
+from langbot.pkg.provider.runners import (
+    cozeapi,
+    dashscopeapi,
+    tboxapi,
+    weknoraapi,
+)
+
+
+@pytest.mark.asyncio
+async def test_blocking_provider_iterator_runs_outside_event_loop():
+    release = threading.Event()
+
+    def values():
+        release.wait(timeout=2)
+        yield 'ready'
+
+    task = asyncio.create_task(anext(runner.iterate_sync(values())))
+    await asyncio.sleep(0)
+    assert not task.done()
+
+    release.set()
+    assert await asyncio.wait_for(task, timeout=1) == 'ready'
+
+
+@pytest.mark.asyncio
+async def test_sync_provider_iterator_has_event_limit():
+    with pytest.raises(RuntimeError, match='event limit'):
+        async for _ in runner.iterate_sync(iter([1, 2]), max_items=1):
+            pass
+
+
+@pytest.mark.asyncio
+async def test_coze_runner_closes_request_scoped_client():
+    request_runner = object.__new__(cozeapi.CozeAPIRunner)
+    request_runner.coze = AsyncMock()
+
+    await request_runner.aclose()
+
+    request_runner.coze.close.assert_awaited_once()
+
+
+@pytest.mark.parametrize(
+    ('append', 'exception_type'),
+    [
+        (cozeapi._append_bounded, ValueError),
+        (dashscopeapi._append_bounded, dashscopeapi.DashscopeAPIError),
+        (tboxapi._append_bounded, tboxapi.TboxAPIError),
+        (weknoraapi._append_bounded, weknoraapi.errors.WeKnoraAPIError),
+    ],
+)
+def test_provider_accumulators_reject_oversized_output(append, exception_type):
+    with pytest.raises(exception_type, match='exceeds the runtime limit'):
+        append('x' * (1024 * 1024), 'y')