backend_skill

Overview

This skill packages Node.js and Firebase backend patterns for scalable, reliable async processing. It focuses on async/await best practices, function sizing, webhook handling, Pub/Sub messaging, cron scheduling, and simple queue patterns. Use it to design systems that avoid cold starts, handle high-volume events, and keep webhooks responsive.

How this skill works

The skill inspects common backend scenarios and recommends concrete patterns: Promise.all for parallel work, chunking for rate-limited bulk operations, and avoiding await inside loops. It explains Firebase Function configuration (memory/timeout), webhook short-response design with background queuing, Pub/Sub publishing/subscribing patterns including fan-out and ordering keys, and alternatives like Cloud Tasks or Firestore queues for delayed or low-volume jobs.

When to use it

• Handling high-volume webhooks or fan-out to multiple consumers
• Building background processing for bulk or long-running jobs
• Designing webhook handlers that must respond quickly (<5s)
• Scheduling recurring work with cron-like functions
• Choosing between Pub/Sub, Cloud Tasks, or Firestore queues based on latency and rate limits

Best practices

• Run independent async calls in parallel with Promise.all; chunk large lists to avoid rate limits
• Never await inside a for-loop; use map+Promise.all or controlled chunking
• Right-size functions: pick memory and timeout based on workload (e.g., 256MB/60s for simple handlers)
• Respond to webhooks immediately and push heavy work to a queue or Pub/Sub
• Prefer Pub/Sub for high volume and fan-out; use Cloud Tasks for delayed or rate-limited external calls
• Configure dead-letter topics and idempotency to handle retries and avoid duplicate side effects

Example use cases

• Receive Shopify webhooks, verify HMAC, publish a single Pub/Sub message and return 200 quickly
• Publish batches of events to a topic with batching for throughput, and let multiple subscribers handle different tasks (points, tiers, notifications)
• Use a Firestore-based queue for low-volume async jobs where setup simplicity matters
• Schedule nightly cleanup and weekly recalculations with scheduled Pub/Sub functions
• Use ordering keys when per-customer sequential processing is required