feat: use workspace identity for telemetry (#2371)

Co-authored-by: dadachann <185672915+dadachann@users.noreply.github.com>

tests/integration/persistence/test_workspace_migration.py

@@ -22,6 +22,7 @@ from langbot.pkg.persistence.alembic_runner import (
 from langbot.pkg.utils import constants
 from langbot.pkg.utils import importutil
 from langbot.pkg.workspace.collaboration import normalize_email
+from langbot.pkg.workspace.identity import workspace_uuid_from_instance_id
 
 
 pytestmark = [pytest.mark.integration, pytest.mark.asyncio]
@@ -109,6 +110,7 @@ async def test_legacy_instance_gets_stable_accounts_and_default_workspace(legacy
             .mappings()
             .one()
         )
+        assert workspace['uuid'] == workspace_uuid_from_instance_id('instance_migration_test')
         assert workspace['instance_uuid'] == 'instance_migration_test'
         assert workspace['slug'] == 'default'
         assert workspace['status'] == 'active'
@@ -149,6 +151,53 @@ async def test_workspace_upgrade_is_idempotent_and_preserves_identifiers(legacy_
     assert workspace_uuid_after == workspace_uuid_before
 
 
+async def test_existing_oss_workspace_is_rekeyed_to_instance_identity(tmp_path):
+    engine = create_async_engine(f'sqlite+aiosqlite:///{tmp_path / "workspace-rekey.db"}')
+    instance_id = 'instance_a711d9e4-0953-443f-a0e9-7dd50193a79f'
+    old_workspace_uuid = '11111111-1111-4111-8111-111111111111'
+    canonical_uuid = workspace_uuid_from_instance_id(instance_id)
+    schema = sa.MetaData()
+    sa.Table(
+        'metadata',
+        schema,
+        sa.Column('key', sa.String(255), primary_key=True),
+        sa.Column('value', sa.String(255)),
+    )
+    sa.Table(
+        'workspaces',
+        schema,
+        sa.Column('uuid', sa.String(36), primary_key=True),
+        sa.Column('instance_uuid', sa.String(255), nullable=False),
+        sa.Column('slug', sa.String(255), nullable=False),
+        sa.Column('source', sa.String(32), nullable=False),
+    )
+    sa.Table(
+        'tenant_rows',
+        schema,
+        sa.Column('id', sa.Integer, primary_key=True),
+        sa.Column('workspace_uuid', sa.String(36), sa.ForeignKey('workspaces.uuid'), nullable=False),
+    )
+    async with engine.begin() as conn:
+        await conn.run_sync(schema.create_all)
+        await conn.execute(sa.text("INSERT INTO metadata (key, value) VALUES ('instance_uuid', :value)"), {'value': instance_id})
+        await conn.execute(
+            sa.text("INSERT INTO workspaces (uuid, instance_uuid, slug, source) VALUES (:uuid, :instance, 'default', 'local')"),
+            {'uuid': old_workspace_uuid, 'instance': instance_id},
+        )
+        await conn.execute(
+            sa.text("INSERT INTO tenant_rows (id, workspace_uuid) VALUES (1, :uuid)"),
+            {'uuid': old_workspace_uuid},
+        )
+    await run_alembic_stamp(engine, '0016_support_admin_sessions')
+
+    await run_alembic_upgrade(engine, 'head')
+
+    async with engine.connect() as conn:
+        assert (await conn.execute(sa.text("SELECT uuid FROM workspaces"))).scalar_one() == canonical_uuid
+        assert (await conn.execute(sa.text("SELECT workspace_uuid FROM tenant_rows"))).scalar_one() == canonical_uuid
+    await engine.dispose()
+
+
 async def test_workspace_kernel_upgrade_downgrade_upgrade_round_trip(tmp_path):
     engine = create_async_engine(f'sqlite+aiosqlite:///{tmp_path / "workspace-round-trip.db"}')
     try:
@@ -362,6 +411,47 @@ async def test_persistence_startup_defers_workspace_tables_until_account_upgrade
         await engine.dispose()
 
 
+async def test_oss_workspace_identity_rekeys_fk_graph_and_metadata(tmp_path):
+    engine = create_async_engine(f'sqlite+aiosqlite:///{tmp_path / "workspace-rekey.db"}')
+    try:
+        await _create_legacy_schema(engine)
+        await run_alembic_stamp(engine, '0008_mcp_resource_prefs')
+        await run_alembic_upgrade(engine, '0016_support_admin_sessions')
+
+        async with engine.begin() as conn:
+            old_uuid = await conn.scalar(sa.text("SELECT uuid FROM workspaces WHERE source = 'local'"))
+            instance_uuid = await conn.scalar(sa.text("SELECT instance_uuid FROM workspaces WHERE source = 'local'"))
+            assert old_uuid
+            assert instance_uuid
+            await conn.execute(
+                sa.text(
+                    "INSERT INTO workspace_metadata (workspace_uuid, key, value) "
+                    "VALUES (:workspace_uuid, 'migration_probe', 'present')"
+                ),
+                {'workspace_uuid': old_uuid},
+            )
+            await conn.execute(
+                sa.text(
+                    "INSERT INTO metadata (key, value) VALUES ('oss_workspace_uuid', :workspace_uuid) "
+                    "ON CONFLICT(key) DO UPDATE SET value = excluded.value"
+                ),
+                {'workspace_uuid': old_uuid},
+            )
+
+        await run_alembic_upgrade(engine, 'head')
+        expected_uuid = workspace_uuid_from_instance_id(instance_uuid)
+        async with engine.connect() as conn:
+            assert await conn.scalar(sa.text("SELECT uuid FROM workspaces WHERE source = 'local'")) == expected_uuid
+            assert await conn.scalar(
+                sa.text("SELECT workspace_uuid FROM workspace_metadata WHERE key = 'migration_probe'")
+            ) == expected_uuid
+            assert await conn.scalar(
+                sa.text("SELECT value FROM metadata WHERE key = 'oss_workspace_uuid'")
+            ) == expected_uuid
+    finally:
+        await engine.dispose()
+
+
 async def test_persistence_startup_rejects_instance_uuid_drift(tmp_path, monkeypatch):
     engine = create_async_engine(f'sqlite+aiosqlite:///{tmp_path / "instance-drift.db"}')
     try:

tests/unit_tests/telemetry/test_heartbeat.py

@@ -60,10 +60,12 @@ class TestBuildHeartbeatPayload:
     async def test_payload_shape(self):
         heartbeat = get_heartbeat_module()
         ap = make_app()
-        payload = await heartbeat.build_heartbeat_payload(ap)
+        payload = await heartbeat.build_heartbeat_payload(ap, workspace_uuid='workspace-a')
 
         assert payload['event_type'] == 'instance_heartbeat'
         assert payload['query_id'] == ''
+        assert payload['workspace_uuid'] == 'workspace-a'
+        assert 'instance_id' not in payload
         assert 'instance_create_ts' in payload
         assert 'timestamp' in payload
         f = payload['features']
@@ -86,7 +88,7 @@ class TestBuildHeartbeatPayload:
     @pytest.mark.asyncio
     async def test_payload_is_json_serializable(self):
         heartbeat = get_heartbeat_module()
-        payload = await heartbeat.build_heartbeat_payload(make_app())
+        payload = await heartbeat.build_heartbeat_payload(make_app(), workspace_uuid='workspace-a')
         json.dumps(payload)
 
     @pytest.mark.asyncio
@@ -94,7 +96,7 @@ class TestBuildHeartbeatPayload:
         heartbeat = get_heartbeat_module()
         ap = make_app()
         ap.persistence_mgr.execute_async = AsyncMock(side_effect=RuntimeError('db down'))
-        payload = await heartbeat.build_heartbeat_payload(ap)
+        payload = await heartbeat.build_heartbeat_payload(ap, workspace_uuid='workspace-a')
         assert payload['features']['pipeline_count'] == -1
 
     @pytest.mark.asyncio
@@ -111,8 +113,10 @@ class TestBuildHeartbeatPayload:
                 ('instance-a', 'workspace-a', 'pipeline-b'): object(),
             },
         )
+        adapter_a = type('WorkspaceAAdapter', (), {})()
+        adapter_b = type('WorkspaceBAdapter', (), {})()
         ap.platform_mgr._bots_by_key = {
-            ('instance-a', 'workspace-a', 'bot-a'): object(),
+            ('instance-a', 'workspace-a', 'bot-a'): SimpleNamespace(enable=True, adapter=adapter_a),
         }
         ap.tool_mgr = SimpleNamespace(
             mcp_tool_loader=SimpleNamespace(
@@ -129,35 +133,46 @@ class TestBuildHeartbeatPayload:
         ap.plugin_connector._workspace_installations = {
             'workspace-a': {'plugin-a', 'plugin-b'},
         }
+        ap.skill_mgr._skills_by_scope = {
+            ('instance-a', 'workspace-a', 1): {'skill-a': {}, 'skill-b': {}},
+            ('instance-a', 'workspace-b', 1): {'skill-c': {}},
+        }
         ap.workspace_service.list_active_execution_bindings = AsyncMock(
-            return_value=[SimpleNamespace(workspace_uuid='workspace-a')],
+            return_value=[
+                SimpleNamespace(workspace_uuid='workspace-a'),
+                SimpleNamespace(workspace_uuid='workspace-b'),
+            ],
+        )
+        ap.platform_mgr._bots_by_key[('instance-a', 'workspace-b', 'bot-b')] = SimpleNamespace(
+            enable=True, adapter=adapter_b
         )
 
-        payload = await heartbeat.build_heartbeat_payload(ap)
+        payloads = await heartbeat.build_heartbeat_payloads(ap)
 
-        features = payload['features']
-        assert features['pipeline_count'] == 2
-        assert features['mcp_server_count'] == 3
-        assert features['knowledge_base_count'] == 1
-        assert features['bot_count'] == 1
-        assert features['workspace_resources'] == [
-            {
-                'workspace_uuid': 'workspace-a',
-                'bot_count': 1,
-                'pipeline_count': 2,
-                'knowledge_base_count': 1,
-                'plugin_count': 2,
-                'mcp_server_count': 3,
-                'extension_count': 5,
-            }
-        ]
+        assert [payload['workspace_uuid'] for payload in payloads] == ['workspace-a', 'workspace-b']
+        assert all('instance_id' not in payload for payload in payloads)
+        by_workspace = {payload['workspace_uuid']: payload['features'] for payload in payloads}
+        assert by_workspace['workspace-a']['pipeline_count'] == 2
+        assert by_workspace['workspace-a']['mcp_server_count'] == 3
+        assert by_workspace['workspace-a']['knowledge_base_count'] == 1
+        assert by_workspace['workspace-a']['bot_count'] == 1
+        assert by_workspace['workspace-a']['plugin_count'] == 2
+        assert by_workspace['workspace-a']['extension_count'] == 5
+        assert by_workspace['workspace-a']['skill_count'] == 2
+        assert by_workspace['workspace-a']['adapters'] == ['WorkspaceAAdapter']
+        assert by_workspace['workspace-b']['bot_count'] == 1
+        assert by_workspace['workspace-b']['pipeline_count'] == 0
+        assert by_workspace['workspace-b']['skill_count'] == 1
+        assert by_workspace['workspace-b']['adapters'] == ['WorkspaceBAdapter']
+        assert 'workspace_resources' not in by_workspace['workspace-a']
         ap.persistence_mgr.execute_async.assert_not_awaited()
+        ap.workspace_service.list_active_execution_bindings.assert_awaited_once()
 
     @pytest.mark.asyncio
     async def test_no_user_content_fields(self):
         """The heartbeat must never carry message content / credentials keys."""
         heartbeat = get_heartbeat_module()
-        payload = await heartbeat.build_heartbeat_payload(make_app())
+        payload = await heartbeat.build_heartbeat_payload(make_app(), workspace_uuid='workspace-a')
         flat = json.dumps(payload).lower()
         for forbidden in ('api_key', 'password', 'token', 'message_content'):
             assert forbidden not in flat

tests/unit_tests/telemetry/test_telemetry.py

@@ -569,6 +569,33 @@ class TestHTTPScenarios:
             await manager.send({'query_id': 'test'})
 
 
+class TestTelemetryManagedRuntimeAuthentication:
+    @pytest.mark.asyncio
+    async def test_send_includes_managed_runtime_token_header(self):
+        telemetry = get_telemetry_module()
+        mock_app = Mock()
+        mock_app.logger = Mock()
+        manager = telemetry.TelemetryManager(mock_app)
+        manager.telemetry_config = {'url': 'https://example.com'}
+        captured = {}
+
+        async def mock_post(url, json, headers):
+            captured['headers'] = headers
+            return Mock(status_code=200, text='', json=Mock(return_value={'code': 0}))
+
+        mock_client = Mock()
+        mock_client.post = mock_post
+        mock_client.__aenter__ = AsyncMock(return_value=mock_client)
+        mock_client.__aexit__ = AsyncMock(return_value=None)
+        with (
+            patch.dict('os.environ', {'LANGBOT_TELEMETRY_INGEST_TOKEN': 'managed-runtime-secret'}),
+            patch.object(httpx, 'AsyncClient', return_value=mock_client),
+        ):
+            await manager.send({'event_type': 'instance_heartbeat'})
+
+        assert captured['headers'] == {'X-LangBot-Telemetry-Token': 'managed-runtime-secret'}
+
+
 class TestStartSendTask:
     """Tests for start_send_task() method."""
 

tests/unit_tests/telemetry/test_workspace_identity.py

@@ -0,0 +1,31 @@
+from __future__ import annotations
+
+import uuid
+from types import SimpleNamespace
+
+
+def test_standard_oss_instance_id_aligns_to_embedded_uuid():
+    from langbot.pkg.workspace.identity import workspace_uuid_from_instance_id
+
+    instance_uuid = "a711d9e4-0953-443f-a0e9-7dd50193a79f"
+
+    assert workspace_uuid_from_instance_id(instance_uuid) == instance_uuid
+    assert workspace_uuid_from_instance_id(f"instance_{instance_uuid}") == instance_uuid
+
+
+def test_custom_legacy_instance_id_maps_to_stable_valid_uuid():
+    from langbot.pkg.workspace.identity import workspace_uuid_from_instance_id
+
+    first = workspace_uuid_from_instance_id("instance_migration_test")
+    second = workspace_uuid_from_instance_id("instance_migration_test")
+
+    assert first == second
+    assert str(uuid.UUID(first)) == first
+
+
+def test_query_telemetry_identity_uses_execution_workspace_only():
+    from langbot.pkg.telemetry.identity import workspace_identity
+
+    identity = workspace_identity(SimpleNamespace(workspace_uuid="workspace-a", instance_uuid="instance-a"))
+
+    assert identity == {"workspace_uuid": "workspace-a"}

tests/unit_tests/workspace/test_workspace_service.py

@@ -26,6 +26,7 @@ from langbot.pkg.workspace import (
     WorkspaceOwnerAlreadyExistsError,
     WorkspaceService,
 )
+from langbot.pkg.workspace.identity import workspace_uuid_from_instance_id
 from langbot.pkg.workspace.policy import CloudWorkspacePolicy
 
 
@@ -113,6 +114,7 @@ async def test_ensure_singleton_workspace_is_idempotent(workspace_test_context):
     first = await service.ensure_singleton_workspace()
     second = await service.ensure_singleton_workspace()
 
+    assert first.uuid == workspace_uuid_from_instance_id('instance_service_test')
     assert second.uuid == first.uuid
     async with session_factory() as session:
         assert await session.scalar(sqlalchemy.select(sqlalchemy.func.count()).select_from(Workspace)) == 1