@vtex/api 7.4.1-beta.1 → 7.4.2-beta.1

package/lib/service/master.d.ts

@@ -1,4 +1,2 @@
-import { Worker } from 'cluster';
 import { ServiceJSON } from './worker/runtime/typings';
-export declare const onMessage: (worker: Worker, message: any) => void;
 export declare const startMaster: (service: ServiceJSON) => void;

package/lib/service/master.js

@@ -3,12 +3,11 @@ var __importDefault = (this && this.__importDefault) || function (mod) {
     return (mod && mod.__esModule) ? mod : { "default": mod };
 };
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.startMaster = exports.onMessage = void 0;
+exports.startMaster = void 0;
 const cluster_1 = __importDefault(require("cluster"));
 const os_1 = require("os");
 const constants_1 = require("../constants");
 const logger_1 = require("./logger");
-const clusterMetricsAggregator_1 = require("./metrics/clusterMetricsAggregator");
 const listeners_1 = require("./worker/listeners");
 const statusTrack_1 = require("./worker/runtime/statusTrack");
 let handledSignal;
@@ -20,12 +19,6 @@ const onMessage = (worker, message) => {
         (0, statusTrack_1.trackStatus)();
         (0, statusTrack_1.broadcastStatusTrack)();
     }
-    else if ((0, clusterMetricsAggregator_1.isAggMetricsRequest)(message)) {
-        (0, clusterMetricsAggregator_1.handleWorkerMetricsRequest)(worker, message);
-    }
-    else if ((0, clusterMetricsAggregator_1.isPromClientMessage)(message)) {
-        // Handled by prom-client's own cluster listener; ignore here.
-    }
     else {
         listeners_1.logger.warn({
             content: message,
@@ -34,7 +27,6 @@ const onMessage = (worker, message) => {
         });
     }
 };
-exports.onMessage = onMessage;
 const onExit = (worker, code, signal) => {
     if (!constants_1.LINKED && worker.exitedAfterDisconnect === false) {
         listeners_1.logger.error({
@@ -78,11 +70,6 @@ const startMaster = (service) => {
     if (service.deterministicVary) {
         process.env.DETERMINISTIC_VARY = 'true';
     }
-    // Set up the master-side Prometheus aggregator so workers can request a
-    // merged, monotonic /metrics view across the whole cluster over IPC.
-    if (numWorkers > 1) {
-        (0, clusterMetricsAggregator_1.initMasterAggregatorRegistry)();
-    }
     // Setup dubugger
     if (constants_1.LINKED) {
         cluster_1.default.setupMaster({ inspectPort: constants_1.INSPECT_DEBUGGER_PORT });
@@ -95,7 +82,7 @@ const startMaster = (service) => {
     }
     cluster_1.default.on('online', onOnline);
     cluster_1.default.on('exit', onExit);
-    cluster_1.default.on('message', exports.onMessage);
+    cluster_1.default.on('message', onMessage);
     process.on('SIGINT', handleSignal(SIGINT_TIMEOUT_S));
     process.on('SIGTERM', handleSignal(shutdownTimeout));
 };

package/lib/service/worker/index.d.ts

@@ -1,4 +1,3 @@
 import Koa from 'koa';
 import { ServiceJSON } from './runtime/typings';
-export declare const onMessage: (service: ServiceJSON) => (message: any) => void;
 export declare const startWorker: (serviceJSON: ServiceJSON) => Koa<Koa.DefaultState, Koa.DefaultContext>;

package/lib/service/worker/index.js

@@ -3,7 +3,7 @@ var __importDefault = (this && this.__importDefault) || function (mod) {
     return (mod && mod.__esModule) ? mod : { "default": mod };
 };
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.startWorker = exports.onMessage = void 0;
+exports.startWorker = void 0;
 const diagnostics_nodejs_1 = require("@vtex/diagnostics-nodejs");
 const http_1 = require("http");
 const koa_1 = __importDefault(require("koa"));
@@ -14,7 +14,6 @@ const constants_1 = require("../../constants");
 const loaders_1 = require("../loaders");
 const console_1 = require("../logger/console");
 const loggerTypes_1 = require("../logger/loggerTypes");
-const clusterMetricsAggregator_1 = require("../metrics/clusterMetricsAggregator");
 const otelRequestMetricsMiddleware_1 = require("../metrics/otelRequestMetricsMiddleware");
 const requestMetricsMiddleware_1 = require("../metrics/requestMetricsMiddleware");
 const TracerSingleton_1 = require("../tracing/TracerSingleton");
@@ -56,12 +55,6 @@ const onMessage = (service) => (message) => {
     else if ((0, statusTrack_1.isStatusTrack)(message)) {
         (0, statusTrack_1.trackStatus)();
     }
-    else if ((0, clusterMetricsAggregator_1.isAggMetricsResponse)(message)) {
-        (0, clusterMetricsAggregator_1.handleMasterMetricsResponse)(message);
-    }
-    else if ((0, clusterMetricsAggregator_1.isPromClientMessage)(message)) {
-        // Handled by prom-client's own worker listener; ignore here.
-    }
     else {
         listeners_1.logger.warn({
             content: message,
@@ -70,7 +63,6 @@ const onMessage = (service) => (message) => {
         });
     }
 };
-exports.onMessage = onMessage;
 const createAppHttpHandlers = ({ config: { routes, clients } }, serviceJSON, globalLimiter) => {
     if (routes && clients) {
         return Object.keys(routes).reduce((acc, routeId) => {
@@ -162,17 +154,12 @@ const startWorker = (serviceJSON) => {
     var _a, _b;
     (0, listeners_1.addProcessListeners)();
     const tracer = TracerSingleton_1.TracerSingleton.getTracer();
-    // In multi-worker mode install prom-client's worker-side cluster responder so
-    // the master can collect this worker's registry for the aggregated /metrics.
-    if (serviceJSON.workers > 1) {
-        (0, clusterMetricsAggregator_1.ensureWorkerAggregatorRegistry)();
-    }
     const app = new koa_1.default();
     app.proxy = true;
     app
         .use(error_1.error)
         .use(diagnostics_nodejs_1.Instrumentation.Middlewares.ContextMiddlewares.Koa.ContextPropagationMiddleware())
-        .use((0, middlewares_1.prometheusLoggerMiddleware)(serviceJSON.workers))
+        .use((0, middlewares_1.prometheusLoggerMiddleware)())
         .use((0, tracingMiddlewares_1.addTracingMiddleware)(tracer))
         .use((0, requestMetricsMiddleware_1.addRequestMetricsMiddleware)())
         .use((0, otelRequestMetricsMiddleware_1.addOtelRequestMetricsMiddleware)())
@@ -197,7 +184,7 @@ const startWorker = (serviceJSON) => {
     ]
         .filter(x => x != null)
         // TODO: Fix ramda typings. Apparently there was an update that broke things
-        .reduce((acc, handler) => (0, ramda_1.mergeDeepRight)(acc, handler), {});
+        .reduce(ramda_1.mergeDeepRight);
     if (httpHandlers === null || httpHandlers === void 0 ? void 0 : httpHandlers.pub) {
         const publicHandlersRouter = (0, router_2.routerFromPublicHttpHandlers)(httpHandlers.pub);
         app.use(publicHandlersRouter);
@@ -206,7 +193,7 @@ const startWorker = (serviceJSON) => {
         const privateHandlersRouter = routerFromPrivateHttpHandlers(httpHandlers.pvt);
         app.use(privateHandlersRouter.routes());
     }
-    process.on('message', (0, exports.onMessage)(serviceJSON));
+    process.on('message', onMessage(serviceJSON));
     return app;
 };
 exports.startWorker = startWorker;

package/lib/service/worker/runtime/builtIn/middlewares.d.ts

@@ -1,4 +1,4 @@
 import { ServiceContext } from '../typings';
 export declare function recorderMiddleware(ctx: ServiceContext, next: () => Promise<void>): Promise<void>;
 export declare const addMetricsLoggerMiddleware: () => (ctx: ServiceContext, next: () => Promise<void>) => Promise<void>;
-export declare const prometheusLoggerMiddleware: (workers?: number) => (ctx: ServiceContext, next: () => Promise<void>) => Promise<void>;
+export declare const prometheusLoggerMiddleware: () => (ctx: ServiceContext, next: () => Promise<void>) => Promise<void>;

package/lib/service/worker/runtime/builtIn/middlewares.js

@@ -4,7 +4,6 @@ exports.prometheusLoggerMiddleware = exports.addMetricsLoggerMiddleware = export
 const prom_client_1 = require("prom-client");
 const constants_1 = require("../../../../constants");
 const metricsLogger_1 = require("../../../logger/metricsLogger");
-const clusterMetricsAggregator_1 = require("../../../metrics/clusterMetricsAggregator");
 const EventLoopLagMeasurer_1 = require("../../../tracing/metrics/measurers/EventLoopLagMeasurer");
 const recorder_1 = require("../utils/recorder");
 async function recorderMiddleware(ctx, next) {
@@ -23,14 +22,10 @@ const addMetricsLoggerMiddleware = () => {
     };
 };
 exports.addMetricsLoggerMiddleware = addMetricsLoggerMiddleware;
-const prometheusLoggerMiddleware = (workers = 1) => {
+const prometheusLoggerMiddleware = () => {
     (0, prom_client_1.collectDefaultMetrics)();
     const eventLoopLagMeasurer = new EventLoopLagMeasurer_1.EventLoopLagMeasurer();
     eventLoopLagMeasurer.start();
-    // In multi-worker mode each worker holds only its own registry, so /metrics
-    // must serve the cluster-wide aggregate requested from the master over IPC.
-    // In single-worker mode (LINKED / workers:1) the local registry is complete.
-    const isMultiWorker = workers > 1;
     return async (ctx, next) => {
         if (ctx.request.path !== '/metrics') {
             return next();
@@ -41,7 +36,7 @@ const prometheusLoggerMiddleware = (workers = 1) => {
         }
         await eventLoopLagMeasurer.updateInstrumentsAndReset();
         ctx.set('Content-Type', prom_client_1.register.contentType);
-        ctx.body = isMultiWorker ? await (0, clusterMetricsAggregator_1.requestAggregatedMetrics)() : await prom_client_1.register.metrics();
+        ctx.body = await prom_client_1.register.metrics();
         ctx.status = 200;
     };
 };

package/lib/service/worker/runtime/statusTrack.js

@@ -26,22 +26,11 @@ exports.statusTrackHandler = statusTrackHandler;
 const trackStatus = () => {
     // Update diagnostics metrics (gauges for HTTP agent stats)
     HttpAgentSingleton_1.HttpAgentSingleton.updateHttpAgentMetrics();
-    // Legacy status tracking (console.log export)
-    global.metrics.statusTrack().forEach(status => {
-        logStatus(status);
-    });
+    // Flushing resets the metric accumulators, the CPU usage baseline and the
+    // incoming request stats, so it must keep running even though nothing
+    // consumes the returned metrics anymore.
+    global.metrics.statusTrack();
 };
 exports.trackStatus = trackStatus;
 const broadcastStatusTrack = () => Object.values(cluster_1.default.workers).forEach(worker => worker === null || worker === void 0 ? void 0 : worker.send(STATUS_TRACK));
 exports.broadcastStatusTrack = broadcastStatusTrack;
-const logStatus = (status) => console.log(JSON.stringify({
-    __VTEX_IO_LOG: true,
-    account: constants_1.ACCOUNT,
-    app: constants_1.APP.ID,
-    isLink: constants_1.LINKED,
-    pid: process.pid,
-    production: constants_1.PRODUCTION,
-    status,
-    type: 'metric/status',
-    workspace: constants_1.WORKSPACE,
-}));

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@vtex/api",
-  "version": "7.4.1-beta.1",
+  "version": "7.4.2-beta.1",
   "description": "VTEX I/O API client",
   "main": "lib/index.js",
   "typings": "lib/index.d.ts",
@@ -132,5 +132,6 @@
   },
   "resolutions": {
     "@grpc/grpc-js": "^1.13.4"
-  }
+  },
+  "packageManager": "yarn@1.22.22+sha512.a6b2f7906b721bba3d67d4aff083df04dad64c399707841b7acf00f6b133b7ac24255f2652fa22ae3534329dc6180534e98d17432037ff6fd140556e2bb3137e"
 }

package/lib/service/metrics/clusterMetricsAggregator.d.ts

@@ -1,72 +0,0 @@
-import { Worker } from 'cluster';
-import { AggregatorRegistry } from 'prom-client';
-/**
- * Cluster-wide Prometheus metrics aggregation.
- *
- * In multi-worker mode each cluster worker keeps its own prom-client default
- * registry. Serving `/metrics` from a single (round-robin selected) worker
- * therefore exposes only that worker's local counters, which makes Prometheus
- * treat the per-scrape braid of independent counters as counter resets and
- * inflates `rate()`/`increase()` by orders of magnitude.
- *
- * This module uses prom-client's `AggregatorRegistry` cluster IPC protocol so
- * that the worker answering a scrape asks the master for a merged, monotonic
- * view built from every worker's registry in one pass (no double-counting of
- * the answering worker).
- */
-/** Tagged IPC message: worker -> master, "please build the cluster aggregate". */
-export declare const AGG_METRICS_REQ = "vtex-api:aggMetricsReq";
-/** Tagged IPC message: master -> worker, carries the aggregate (or an error). */
-export declare const AGG_METRICS_RES = "vtex-api:aggMetricsRes";
-export interface AggMetricsReqMessage {
-    type: typeof AGG_METRICS_REQ;
-    id: number;
-}
-export interface AggMetricsResMessage {
-    type: typeof AGG_METRICS_RES;
-    id: number;
-    body?: string;
-    error?: string;
-}
-export declare const isAggMetricsRequest: (message: any) => message is AggMetricsReqMessage;
-export declare const isAggMetricsResponse: (message: any) => message is AggMetricsResMessage;
-/**
- * prom-client's cluster protocol emits its own tagged IPC messages
- * (`prom-client:getMetricsReq` / `prom-client:getMetricsRes`). They are handled
- * by prom-client's own cluster listeners, so our `onMessage` handlers must
- * ignore them instead of warning about "unknown" messages.
- */
-export declare const isPromClientMessage: (message: any) => boolean;
-/**
- * Constructs the master-side `AggregatorRegistry` (idempotent). Its constructor
- * installs prom-client's master `cluster.on('message')` collector, which is
- * what makes `clusterMetrics()` work. Must be called in the master before any
- * worker can answer a scrape.
- */
-export declare const initMasterAggregatorRegistry: () => AggregatorRegistry;
-/**
- * Handles an `AGG_METRICS_REQ` from a worker: builds the cluster aggregate and
- * replies to that worker only. prom-client fans the request out to *all*
- * connected workers (including the requester), so the requester is one of the N
- * merged sources and is never double-counted.
- */
-export declare const handleWorkerMetricsRequest: (worker: Worker, message: AggMetricsReqMessage) => Promise<void>;
-/**
- * Constructs the worker-side `AggregatorRegistry` (idempotent). Its constructor
- * installs prom-client's worker `process.on('message')` responder that answers
- * the master's `getMetricsReq` with this worker's registry JSON.
- */
-export declare const ensureWorkerAggregatorRegistry: () => AggregatorRegistry;
-/**
- * Requests the cluster aggregate from the master over IPC and resolves with the
- * merged exposition string. On timeout or any error it falls back to this
- * worker's local `register.metrics()` so `/metrics` always answers promptly.
- */
-export declare const requestAggregatedMetrics: () => Promise<string>;
-/**
- * Resolves the pending `requestAggregatedMetrics` promise correlated by id when
- * the master replies. On error it falls back to the local registry.
- */
-export declare const handleMasterMetricsResponse: (message: AggMetricsResMessage) => void;
-/** Test-only helper to reset module state between test cases. */
-export declare const __resetForTests: () => void;

package/lib/service/metrics/clusterMetricsAggregator.js

@@ -1,164 +0,0 @@
-"use strict";
-Object.defineProperty(exports, "__esModule", { value: true });
-exports.__resetForTests = exports.handleMasterMetricsResponse = exports.requestAggregatedMetrics = exports.ensureWorkerAggregatorRegistry = exports.handleWorkerMetricsRequest = exports.initMasterAggregatorRegistry = exports.isPromClientMessage = exports.isAggMetricsResponse = exports.isAggMetricsRequest = exports.AGG_METRICS_RES = exports.AGG_METRICS_REQ = void 0;
-const prom_client_1 = require("prom-client");
-const listeners_1 = require("../worker/listeners");
-/**
- * Cluster-wide Prometheus metrics aggregation.
- *
- * In multi-worker mode each cluster worker keeps its own prom-client default
- * registry. Serving `/metrics` from a single (round-robin selected) worker
- * therefore exposes only that worker's local counters, which makes Prometheus
- * treat the per-scrape braid of independent counters as counter resets and
- * inflates `rate()`/`increase()` by orders of magnitude.
- *
- * This module uses prom-client's `AggregatorRegistry` cluster IPC protocol so
- * that the worker answering a scrape asks the master for a merged, monotonic
- * view built from every worker's registry in one pass (no double-counting of
- * the answering worker).
- */
-/** Tagged IPC message: worker -> master, "please build the cluster aggregate". */
-exports.AGG_METRICS_REQ = 'vtex-api:aggMetricsReq';
-/** Tagged IPC message: master -> worker, carries the aggregate (or an error). */
-exports.AGG_METRICS_RES = 'vtex-api:aggMetricsRes';
-/**
- * Upper bound for a worker waiting on the master aggregate. prom-client's own
- * `clusterMetrics()` has a 5s timeout collecting worker registries; we guard
- * slightly above that and fall back to the local registry on expiry so that
- * `/metrics` never hangs or returns empty.
- */
-const AGG_METRICS_TIMEOUT_MS = 6000;
-const isAggMetricsRequest = (message) => (message === null || message === void 0 ? void 0 : message.type) === exports.AGG_METRICS_REQ && typeof (message === null || message === void 0 ? void 0 : message.id) === 'number';
-exports.isAggMetricsRequest = isAggMetricsRequest;
-const isAggMetricsResponse = (message) => (message === null || message === void 0 ? void 0 : message.type) === exports.AGG_METRICS_RES && typeof (message === null || message === void 0 ? void 0 : message.id) === 'number';
-exports.isAggMetricsResponse = isAggMetricsResponse;
-/**
- * prom-client's cluster protocol emits its own tagged IPC messages
- * (`prom-client:getMetricsReq` / `prom-client:getMetricsRes`). They are handled
- * by prom-client's own cluster listeners, so our `onMessage` handlers must
- * ignore them instead of warning about "unknown" messages.
- */
-const isPromClientMessage = (message) => typeof (message === null || message === void 0 ? void 0 : message.type) === 'string' && message.type.startsWith('prom-client:');
-exports.isPromClientMessage = isPromClientMessage;
-// ---------------------------------------------------------------------------
-// Master side
-// ---------------------------------------------------------------------------
-let masterAggregatorRegistry;
-/**
- * Constructs the master-side `AggregatorRegistry` (idempotent). Its constructor
- * installs prom-client's master `cluster.on('message')` collector, which is
- * what makes `clusterMetrics()` work. Must be called in the master before any
- * worker can answer a scrape.
- */
-const initMasterAggregatorRegistry = () => {
-    if (!masterAggregatorRegistry) {
-        masterAggregatorRegistry = new prom_client_1.AggregatorRegistry();
-    }
-    return masterAggregatorRegistry;
-};
-exports.initMasterAggregatorRegistry = initMasterAggregatorRegistry;
-/**
- * Handles an `AGG_METRICS_REQ` from a worker: builds the cluster aggregate and
- * replies to that worker only. prom-client fans the request out to *all*
- * connected workers (including the requester), so the requester is one of the N
- * merged sources and is never double-counted.
- */
-const handleWorkerMetricsRequest = async (worker, message) => {
-    var _a;
-    const aggregator = (0, exports.initMasterAggregatorRegistry)();
-    try {
-        const body = await aggregator.clusterMetrics();
-        worker.send({ type: exports.AGG_METRICS_RES, id: message.id, body });
-    }
-    catch (err) {
-        worker.send({ type: exports.AGG_METRICS_RES, id: message.id, error: (_a = err === null || err === void 0 ? void 0 : err.message) !== null && _a !== void 0 ? _a : String(err) });
-    }
-};
-exports.handleWorkerMetricsRequest = handleWorkerMetricsRequest;
-// ---------------------------------------------------------------------------
-// Worker side
-// ---------------------------------------------------------------------------
-let workerAggregatorRegistry;
-const pendingRequests = new Map();
-let requestCounter = 0;
-/**
- * Constructs the worker-side `AggregatorRegistry` (idempotent). Its constructor
- * installs prom-client's worker `process.on('message')` responder that answers
- * the master's `getMetricsReq` with this worker's registry JSON.
- */
-const ensureWorkerAggregatorRegistry = () => {
-    if (!workerAggregatorRegistry) {
-        workerAggregatorRegistry = new prom_client_1.AggregatorRegistry();
-    }
-    return workerAggregatorRegistry;
-};
-exports.ensureWorkerAggregatorRegistry = ensureWorkerAggregatorRegistry;
-/**
- * Requests the cluster aggregate from the master over IPC and resolves with the
- * merged exposition string. On timeout or any error it falls back to this
- * worker's local `register.metrics()` so `/metrics` always answers promptly.
- */
-const requestAggregatedMetrics = async () => {
-    if (typeof process.send !== 'function') {
-        return prom_client_1.register.metrics();
-    }
-    const id = requestCounter++;
-    return new Promise((resolve) => {
-        var _a;
-        const timer = setTimeout(() => {
-            pendingRequests.delete(id);
-            listeners_1.logger.warn({
-                message: 'Timed out waiting for aggregated cluster metrics; falling back to local registry',
-                pid: process.pid,
-            });
-            prom_client_1.register.metrics().then(resolve).catch(() => resolve(''));
-        }, AGG_METRICS_TIMEOUT_MS);
-        pendingRequests.set(id, { resolve, timer });
-        try {
-            process.send({ type: exports.AGG_METRICS_REQ, id });
-        }
-        catch (err) {
-            clearTimeout(timer);
-            pendingRequests.delete(id);
-            listeners_1.logger.warn({
-                content: (_a = err === null || err === void 0 ? void 0 : err.message) !== null && _a !== void 0 ? _a : String(err),
-                message: 'Failed to request aggregated cluster metrics; falling back to local registry',
-                pid: process.pid,
-            });
-            prom_client_1.register.metrics().then(resolve).catch(() => resolve(''));
-        }
-    });
-};
-exports.requestAggregatedMetrics = requestAggregatedMetrics;
-/**
- * Resolves the pending `requestAggregatedMetrics` promise correlated by id when
- * the master replies. On error it falls back to the local registry.
- */
-const handleMasterMetricsResponse = (message) => {
-    const pending = pendingRequests.get(message.id);
-    if (!pending) {
-        return;
-    }
-    pendingRequests.delete(message.id);
-    clearTimeout(pending.timer);
-    if (message.error != null || message.body == null) {
-        listeners_1.logger.warn({
-            content: message.error,
-            message: 'Master failed to aggregate cluster metrics; falling back to local registry',
-            pid: process.pid,
-        });
-        prom_client_1.register.metrics().then(pending.resolve).catch(() => pending.resolve(''));
-        return;
-    }
-    pending.resolve(message.body);
-};
-exports.handleMasterMetricsResponse = handleMasterMetricsResponse;
-/** Test-only helper to reset module state between test cases. */
-const __resetForTests = () => {
-    masterAggregatorRegistry = undefined;
-    workerAggregatorRegistry = undefined;
-    pendingRequests.forEach((p) => clearTimeout(p.timer));
-    pendingRequests.clear();
-    requestCounter = 0;
-};
-exports.__resetForTests = __resetForTests;