---
name: throttling-apis
description: 'Implement API throttling policies to protect backend services from overload.

  Use when controlling API request rates.

  Trigger with phrases like "throttle API", "control request rate", or "add throttling".'
allowed-tools: Read, Write, Edit, Grep, Glob, Bash(api:throttle-*)
version: 1.26.0
author: Jeremy Longshore <jeremy@intentsolutions.io>
license: MIT
tags:
  - api
  - throttling-apis
compatibility: Designed for Claude Code
title: throttling-apis
canonical_url: https://skilld.dev/gh/jeremylongshore/tons-of-skills-marketplace/throttling-apis
last_updated: 2026-10-23T05:08:30.000Z
---

> Index of every page on this site: [llms.txt](https://skilld.dev/llms.txt)

> **Skill from skilld.dev.** Follow the user's requested action.
> If the user asked to fork this Skill, follow the fork workflow below. Do not execute the Skill instructions while copying it.
> Otherwise, follow the instructions below for this session. You do not need to install anything.
>
> Supporting files, fetch one when the Skill refers to it: [references/errors.md](https://skilld.dev/api/skills-raw/jeremylongshore/tons-of-skills-marketplace/throttling-apis/references/errors.md), [references/examples.md](https://skilld.dev/api/skills-raw/jeremylongshore/tons-of-skills-marketplace/throttling-apis/references/examples.md), [references/implementation.md](https://skilld.dev/api/skills-raw/jeremylongshore/tons-of-skills-marketplace/throttling-apis/references/implementation.md).
>
> If the user asked to install this Skill, run `npx skilld install jeremylongshore/tons-of-skills-marketplace/throttling-apis`. Install writes the Skill files into the project, so every session loads them.
>
> ## Fork workflow
>
> A fork creates an editable local Skill with its original author and licence. The request authorizes copying and local installation.
> 1. Check `./skills/throttling-apis`, the project lockfile, and selected Agent targets together. If the local directory or installed Skill exists, stop. Never overwrite an existing directory or Agent target.
> 2. Read [source metadata](https://skilld.dev/api/v1/skills/jeremylongshore/tons-of-skills-marketplace/throttling-apis) once. Use sourceUrl, sourceCommit, skillPath, sourceGone, and license. If the source is gone or its path is missing, stop. If license is null, read licence files at the source commit.
> 3. Fetch only the source commit into a temporary Git repository. Do not clone full history. Derive repository_url from sourceUrl, including repository renames. If sourceCommit is absent, resolve the sourceUrl ref once. Set source_commit to that actual commit. Run these commands in one shell call:
>
> ```sh
> git init --quiet "$temporary_dir"
> git -C "$temporary_dir" fetch --quiet --depth=1 "$repository_url" "$source_commit"
> git -C "$temporary_dir" checkout --quiet --detach FETCH_HEAD
> ```
>
> Read applicable licence declarations and notices at that commit. If copying is not permitted, report the restriction and stop.
> 4. Inspect source entries together, then copy the directory containing skillPath into `./skills/throttling-apis`. Use the user's path if selected. Keep the original SKILL.md, relative links, scripts, binary assets, and executable modes. Exclude .git metadata. Reject symlinks and paths outside the Skill directory. After checking entries, use cp -a where available. A regular source directory needs no custom copy script. Do not save this page wrapper as SKILL.md.
> Preserve author credit, notices, and applicable licence files from repository or parent directories. Add PROVENANCE.md with the Skill page, source URL, actual commit, original path, and licence. Retain any existing PROVENANCE.md and record new provenance separately. Batch source inspection, copying, and provenance work where practical.
> 5. In the project root, run `skilld install ./skills/throttling-apis --mode copy --plain`. If skilld is unavailable, use `npx skilld install ./skills/throttling-apis --mode copy --plain`. This known command needs no help lookup. Install does not support --json. Use detected Agent targets, or add --agent for the targets the user selected. Install the local path, never the upstream selector. If installation fails, preserve the local copy and report the exact failure.
> 6. Confirm the local lockfile source and installed Agent copies once. Report the local path, actual commit, and Agent targets. After edits, reinstall the same local path. Upstream updates must not replace it. Do not publish or push unless the user asks.

# Throttling APIs

## Overview

Implement API throttling policies that protect backend services from overload by controlling request concurrency, queue depth, and processing rates. Apply backpressure mechanisms including concurrent request limits, priority queues, circuit breakers, and adaptive throttling that adjusts limits based on real-time backend health metrics.

## Prerequisites

- Middleware-capable web framework (Express, FastAPI, Spring Boot, Gin)
- Redis or in-memory store for distributed throttle state tracking
- Monitoring system exposing backend latency and error rate metrics (Prometheus, CloudWatch)
- Load testing tool (k6, Artillery, wrk) for validating throttle behavior under pressure
- Queue system for request buffering during throttle events (optional: Bull, SQS)

## Instructions

1. Analyze existing route handlers and middleware using Grep and Read to identify endpoints with high latency, database-heavy operations, or external service dependencies that need throttle protection.
2. Implement a concurrency limiter middleware that tracks in-flight requests per endpoint and rejects new requests with 503 Service Unavailable when the concurrent limit is reached.
3. Add priority queue support that classifies requests by API key tier (free, pro, enterprise) and serves higher-tier requests first when approaching throttle limits.
4. Build a circuit breaker for downstream service calls that opens after configurable failure thresholds (e.g., 5 failures in 10 seconds), returning 503 with `Retry-After` during the open state.
5. Configure adaptive throttling that monitors backend response latency percentiles (p95, p99) and automatically reduces concurrency limits when latency exceeds SLO thresholds.
6. Add throttle state headers to all responses: `X-Throttle-Limit`, `X-Throttle-Remaining`, and `X-Throttle-Reset` for client-side awareness.
7. Implement graceful degradation strategies per endpoint: serve cached responses, return partial results, or queue requests for deferred processing.
8. Write load tests that verify throttle engagement at expected thresholds, proper 503 responses with `Retry-After`, and recovery behavior when load subsides.

See [`references/implementation.md`](https://skilld.dev/api/skills-raw/jeremylongshore/tons-of-skills-marketplace/throttling-apis/references/implementation.md) for the full implementation guide.

## Output

- `${CLAUDE_SKILL_DIR}/src/middleware/throttle.js` - Concurrency and request rate throttling middleware
- `${CLAUDE_SKILL_DIR}/src/middleware/circuit-breaker.js` - Circuit breaker for downstream service protection
- `${CLAUDE_SKILL_DIR}/src/middleware/priority-queue.js` - Tier-based request prioritization
- `${CLAUDE_SKILL_DIR}/src/config/throttle-config.js` - Per-endpoint throttle policy definitions
- `${CLAUDE_SKILL_DIR}/tests/throttle/` - Load tests validating throttle engagement and recovery

## Error Handling

| Error | Cause | Solution |
|-------|-------|----------|
| 503 Service Unavailable | Concurrency limit reached for the endpoint | Return `Retry-After` header with estimated wait time; include throttle state headers |
| 503 Circuit Open | Circuit breaker tripped due to downstream failures | Return cached response if available; provide circuit reset time in response body |
| Queue overflow | Request buffer exceeded maximum depth | Reject with 503; alert operations team; consider scaling backend capacity |
| Stale throttle state | Redis connection lost; throttle counters become inaccurate | Fall back to in-process counters; reconnect with backoff; log state inconsistency |
| Priority starvation | Low-tier requests never served under sustained high-tier load | Reserve minimum throughput percentage for each tier to prevent complete starvation |

Refer to [`references/errors.md`](https://skilld.dev/api/skills-raw/jeremylongshore/tons-of-skills-marketplace/throttling-apis/references/errors.md) for comprehensive error patterns.

## Examples

**Database-heavy endpoint protection**: Apply concurrency limit of 10 to a report generation endpoint that runs expensive aggregation queries, queueing additional requests with estimated wait times.

**Multi-tier SaaS throttling**: Enterprise tier gets 100 concurrent requests, Pro tier gets 25, Free tier gets 5, with priority queue ensuring enterprise requests are served first during contention.

**Adaptive autoscaling trigger**: Throttle middleware emits metrics that trigger horizontal pod autoscaling when throttle engagement rate exceeds 20% sustained over 5 minutes.

See [`references/examples.md`](https://skilld.dev/api/skills-raw/jeremylongshore/tons-of-skills-marketplace/throttling-apis/references/examples.md) for additional examples.

## Resources

- Circuit Breaker pattern: Martin Fowler's design patterns
- Resilience4j (Java) and cockatiel (Node.js) circuit breaker libraries
- Netflix Concurrency Limits library for adaptive throttling
- Token bucket and leaky bucket algorithm implementations
