agentleFS
roedyrustamroedyrustam/vibes-plug/skills/async-queue-temporal-expert/SKILL.md

async-queue-temporal-expert

Skill75 starsChanged 38 days agoMITView on GitHub

Unified expert guide for async job queues & durable workflows: BullMQ v5 (Redis queues), Trigger.dev v3 (serverless tasks), Inngest, and Temporal.io (distributed sagas) / Panduan ahli terpadu untuk antrean job asinkron & workflow tahan-gagal: BullMQ v5, Trigger.dev v3, Inngest, dan Temporal.io.

  • Reads credentials

The file

Files
SKILL.md

Async Queue & Durable Workflow Expert (2026 Unified Edition)

English | Bahasa Indonesia


<a name="english"></a>

English

Purpose & Overview

Unified production-grade guide for background execution pipelines, async job queues, and distributed state machines. Covers the full spectrum from simple Redis-backed task queues to complex multi-service sagas with compensating rollbacks.

3-Tier Execution Model

TierEngineBest For
Tier 1: Redis Task QueuesBullMQ v5High-throughput worker jobs, priority queues, rate limiting, DLQ
Tier 2: Serverless Durable TasksTrigger.dev v3 / InngestStep-checkpointed tasks, automatic resume across crashes, zero infra
Tier 3: Distributed SagasTemporal.ioMulti-service orchestration, compensating rollbacks, long-running workflows

Core Capabilities

  1. Idempotency & Deduplication: Deterministic jobId keys prevent double billing or duplicate emails.
  2. Dead Letter Queues (DLQ): Auto-relocate permanently failing jobs for audit and alerting.
  3. Exponential Backoff with Jitter: Prevents thundering herds on upstream services.
  4. Tenant Priority Queues: VIP/enterprise tiers get lower BullMQ priority numbers (higher throughput).
  5. Durable State Machines: Workflows survive restarts, deployments, and network partitions.
  6. Saga Compensations: Multi-step transactions with automated reverse-order rollbacks.

Tier 1: BullMQ v5 — Redis Task Queues (TypeScript)

import { Queue, Worker, Job } from 'bullmq';
import Redis from 'ioredis';

const redisConnection = new Redis(process.env.REDIS_URL!, {
  maxRetriesPerRequest: null, // Required by BullMQ
});

export interface NotificationPayload {
  tenantId: string;
  userId: string;
  type: 'email' | 'webhook';
  payload: Record<string, unknown>;
  idempotencyKey: string;
}

// Main Queue
export const notificationQueue = new Queue<NotificationPayload>('notifications', {
  connection: redisConnection,
  defaultJobOptions: {
    attempts: 5,
    backoff: { type: 'exponential', delay: 1500 },
    removeOnComplete: { age: 86400, count: 5000 },
    removeOnFail: false, // Preserved for DLQ audit
  },
});

// Dead Letter Queue
export const notificationDLQ = new Queue('notifications-dlq', {
  connection: redisConnection,
});

// Enqueue with deduplication & priority
export async function enqueueNotification(data: NotificationPayload, isVip = false) {
  return await notificationQueue.add('send_notification', data, {
    jobId: `notif_${data.idempotencyKey}`, // Deterministic dedup key
    priority: isVip ? 1 : 10,
  });
}

// Worker with concurrency & rate limiting
export const notificationWorker = new Worker<NotificationPayload>(
  'notifications',
  async (job: Job<NotificationPayload>) => {
    if (job.data.type === 'email') await deliverEmail(job.data);
  },
  {
    connection: redisConnection,
    concurrency: 20,
    limiter: { max: 100, duration: 1000 },
  }
);

// DLQ forwarding on retry exhaustion
notificationWorker.on('failed', async (job, error) => {
  if (job && job.attemptsMade >= (job.opts.attempts || 5)) {
    await notificationDLQ.add('failed_notification', {
      originalJobId: job.id, failedReason: error.message,
      data: job.data, exhaustedAt: new Date().toISOString(),
    });
  }
});

Tier 2: Trigger.dev v3 — Serverless Durable Tasks

import { task } from '@trigger.dev/sdk/v3';

export const generateReport = task({
  id: 'generate-enterprise-report',
  retry: { maxAttempts: 4, minTimeoutInMs: 2000, factor: 2, randomize: true },
  run: async (payload: { tenantId: string; month: string }, { ctx }) => {
    // Each step is a durable checkpoint — survives crashes
    const data = await ctx.run('fetch-telemetry', async () => {
      return await fetchTelemetryFromWarehouse(payload.tenantId, payload.month);
    });
    const pdfUrl = await ctx.run('render-pdf', async () => {
      return await generateReportPdf(data);
    });
    await ctx.run('dispatch-webhook', async () => {
      return await sendWebhookNotification(payload.tenantId, pdfUrl);
    });
    return { success: true, pdfUrl };
  },
});

Tier 3: Temporal.io — Distributed Saga with Compensations

import { proxyActivities, ApplicationFailure } from '@temporalio/workflow';
import type * as activities from './activities';

const { chargeCustomer, provisionLicense, sendWelcomeEmail, refundCustomer, revokeLicense } =
  proxyActivities<typeof activities>({
    startToCloseTimeout: '1 minute',
    retry: {
      initialInterval: '1s', backoffCoefficient: 2, maximumAttempts: 5,
      nonRetryableErrorTypes: ['InvalidCardError', 'AccountSuspendedError'],
    },
  });

export async function subscriptionSagaWorkflow(input: {
  customerId: string; planId: string; amountCents: number;
}) {
  const compensations: Array<() => Promise<void>> = [];
  try {
    const charge = await chargeCustomer(input.customerId, input.amountCents);
    compensations.unshift(() => refundCustomer(charge.chargeId));

    const license = await provisionLicense(input.customerId, input.planId);
    compensations.unshift(() => revokeLicense(license.licenseId));

    await sendWelcomeEmail(input.customerId, license.licenseKey);
    return { status: 'COMPLETED' };
  } catch (error) {
    for (const compensate of compensations) {
      try { await compensate(); } catch (e) { console.error('Compensation failed:', e); }
    }
    throw ApplicationFailure.create({
      message: `Saga rolled back: ${(error as Error).message}`, nonRetryable: true,
    });
  }
}

Temporal Determinism Rule: Never use Math.random(), Date.now(), or direct DB calls inside workflow files. Run them inside activities.


Implementation Checklist

  • Configure maxRetriesPerRequest: null on ioredis for BullMQ v5.
  • Use deterministic jobId from business logic (order_${orderId}) for deduplication.
  • Forward permanently dead jobs to DLQ via worker.on('failed') listener.
  • Implement rate limiting via worker limiter to protect third-party APIs.
  • Enforce deterministic code inside Temporal workflows (activities for side-effects).
  • Store large payloads in S3/R2; pass only IDs through queues.
  • Implement Saga rollback handlers for multi-step distributed payments.

Orchestration & Integration

  • Integrates with: js-backend-expert, cron-scheduler-expert, error-resilience-expert, saas-billing, doku-payment-gateway, data-telemetry-expert.

<a name="bahasa-indonesia"></a>

Bahasa Indonesia

Tujuan & Gambaran Umum

Panduan terpadu tingkat produksi untuk pipeline eksekusi latar belakang, antrean job asinkron, dan state machine terdistribusi. Mencakup spektrum penuh dari antrean Redis sederhana hingga saga multi-layanan dengan rollback kompensasi.

Model Eksekusi 3-Tier

TierEngineCocok Untuk
Tier 1: Antrean RedisBullMQ v5Job worker throughput tinggi, prioritas, rate limiting, DLQ
Tier 2: Task ServerlessTrigger.dev v3 / InngestTask dengan checkpoint, resume otomatis, tanpa infra
Tier 3: Saga TerdistribusiTemporal.ioOrkestrasi multi-layanan, rollback kompensasi, workflow jangka panjang

Kemampuan Utama

  1. Idempotensi & Deduplikasi: Kunci jobId deterministik mencegah duplikasi penagihan atau email.
  2. Dead Letter Queue (DLQ): Pemindahan otomatis job gagal total untuk audit.
  3. Backoff Eksponensial + Jitter: Mencegah thundering herd pada server hilir.
  4. Prioritas Tenant: Tier VIP/enterprise mendapat prioritas lebih tinggi (angka lebih kecil di BullMQ).
  5. State Machine Tahan-Gagal: Workflow bertahan saat restart, deployment, dan partisi jaringan.
  6. Kompensasi Saga: Transaksi multi-langkah dengan rollback otomatis urutan mundur.

Checklist Implementasi

  • Atur maxRetriesPerRequest: null pada ioredis untuk BullMQ v5.
  • Gunakan jobId deterministik dari ID bisnis (invoice_${invoiceId}) untuk deduplikasi.
  • Pasang listener worker.on('failed') untuk forward job gagal ke DLQ.
  • Terapkan rate limiter pada worker untuk stabilitas API eksternal.
  • Pastikan kode Temporal selalu deterministik (side-effect hanya di activities).
  • Simpan file besar di S3/R2; kirim hanya referensi ID melalui queue.

Integrasi Orkestrasi

  • Terintegrasi dengan: js-backend-expert, cron-scheduler-expert, error-resilience-expert, saas-billing, doku-payment-gateway, data-telemetry-expert.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.