fix(elastic-console): expose model context window sizes

Surface connector context window metadata from the inference plugin so OpenAI-compatible clients can size prompts per model.

Co-authored-by: Cursor <cursoragent@cursor.com>

Author: vigneshshanmugam

x-pack/platform/plugins/shared/elastic_console/jest.config.js

@@ -0,0 +1,16 @@
+/*
+ * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
+ * or more contributor license agreements. Licensed under the Elastic License
+ * 2.0; you may not use this file except in compliance with the Elastic License
+ * 2.0.
+ */
+
+module.exports = {
+  preset: '@kbn/test',
+  rootDir: '../../../../..',
+  roots: [
+    '<rootDir>/x-pack/platform/plugins/shared/elastic_console/common',
+    '<rootDir>/x-pack/platform/plugins/shared/elastic_console/public',
+    '<rootDir>/x-pack/platform/plugins/shared/elastic_console/server',
+  ],
+};

x-pack/platform/plugins/shared/elastic_console/server/routes/models.test.ts

@@ -0,0 +1,117 @@
+/*
+ * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
+ * or more contributor license agreements. Licensed under the Elastic License
+ * 2.0; you may not use this file except in compliance with the Elastic License
+ * 2.0.
+ */
+
+import type { CoreSetup } from '@kbn/core/server';
+import { httpServerMock, httpServiceMock, loggingSystemMock } from '@kbn/core/server/mocks';
+import { type InferenceConnector, InferenceConnectorType } from '@kbn/inference-common';
+import { inferenceMock } from '@kbn/inference-plugin/server/mocks';
+import type { ElasticConsolePluginStart, ElasticConsoleStartDependencies } from '../types';
+import { isElasticConsoleEnabled } from './is_enabled';
+import { registerModelsRoute } from './models';
+
+jest.mock('./is_enabled', () => ({
+  isElasticConsoleEnabled: jest.fn(),
+}));
+
+const isElasticConsoleEnabledMock = isElasticConsoleEnabled as jest.MockedFunction<
+  typeof isElasticConsoleEnabled
+>;
+
+const createConnector = (parts: Partial<InferenceConnector> = {}): InferenceConnector => ({
+  connectorId: 'connector-1',
+  name: 'Connector 1',
+  type: InferenceConnectorType.OpenAI,
+  config: {},
+  capabilities: {},
+  isInferenceEndpoint: false,
+  isPreconfigured: false,
+  ...parts,
+});
+
+describe('registerModelsRoute', () => {
+  let coreSetup: jest.Mocked<
+    Pick<CoreSetup<ElasticConsoleStartDependencies, ElasticConsolePluginStart>, 'getStartServices'>
+  >;
+  let inference: ReturnType<typeof inferenceMock.createStartContract>;
+  let router: ReturnType<typeof httpServiceMock.createRouter>;
+  const logger = loggingSystemMock.create().get();
+
+  beforeEach(() => {
+    jest.spyOn(Date, 'now').mockReturnValue(1700000000000);
+
+    inference = inferenceMock.createStartContract();
+    router = httpServiceMock.createRouter();
+    coreSetup = {
+      getStartServices: jest.fn().mockResolvedValue([{}, { inference }]),
+    };
+    isElasticConsoleEnabledMock.mockResolvedValue(true);
+  });
+
+  afterEach(() => {
+    jest.restoreAllMocks();
+  });
+
+  it('registers the OpenAI-compatible models route', () => {
+    registerModelsRoute({ router, coreSetup: coreSetup as CoreSetup, logger });
+
+    const [config] = router.get.mock.calls[0];
+    expect(config.path).toBe('/internal/elastic_ramen/v1/models');
+    expect(config.options?.access).toBe('internal');
+  });
+
+  it('returns connector context window sizes when available', async () => {
+    const request = httpServerMock.createKibanaRequest();
+    const response = httpServerMock.createResponseFactory();
+    inference.getConnectorList.mockResolvedValue([
+      createConnector({
+        connectorId: 'sonnet-4',
+        type: InferenceConnectorType.Bedrock,
+        capabilities: { contextWindowSize: 1000000 },
+      }),
+      createConnector({ connectorId: 'unknown-model' }),
+    ]);
+
+    registerModelsRoute({ router, coreSetup: coreSetup as CoreSetup, logger });
+
+    const [, handler] = router.get.mock.calls[0];
+    await handler({}, request, response);
+
+    expect(inference.getConnectorList).toHaveBeenCalledWith(request);
+    expect(response.ok).toHaveBeenCalledWith({
+      body: {
+        object: 'list',
+        data: [
+          {
+            id: 'sonnet-4',
+            object: 'model',
+            created: 1700000000,
+            owned_by: InferenceConnectorType.Bedrock,
+            permission: [],
+            root: 'sonnet-4',
+            parent: null,
+            context_window_size: 1000000,
+          },
+          {
+            id: 'unknown-model',
+            object: 'model',
+            created: 1700000000,
+            owned_by: InferenceConnectorType.OpenAI,
+            permission: [],
+            root: 'unknown-model',
+            parent: null,
+            context_window_size: undefined,
+          },
+        ],
+      },
+    });
+
+    const [{ body }] = response.ok.mock.calls[0];
+    const serializedBody = JSON.parse(JSON.stringify(body));
+    expect(serializedBody.data[0].context_window_size).toBe(1000000);
+    expect(serializedBody.data[1]).not.toHaveProperty('context_window_size');
+  });
+});

x-pack/platform/plugins/shared/elastic_console/server/routes/models.ts

@@ -47,6 +47,7 @@ export const registerModelsRoute = ({
           permission: [],
           root: connector.connectorId,
           parent: null,
+          context_window_size: connector.capabilities.contextWindowSize,
         }));
 
         return response.ok({