Protect public webhooks with Ainoflow Guard rate limiting

Webhook Rate Limiter (Ainoflow Guard)

Stop webhook flooding before it starts. Add production-grade rate limiting to any n8n webhook in minutes - reject abusive traffic before expensive workflow logic executes.

✨ Key Features

• ⚡ Edge-style decisions - Allow/deny checked before any business logic runs
• 🛡️ Burst protection - Configurable limits (requests per time window)
• 🔄 Stateless - No queues, databases, or counters needed in n8n
• 📡 Proxy-aware - Correct IP extraction behind Cloudflare, nginx, load balancers
• 🔑 Dual identity modes - Rate limit by IP address or API key
• ⏱️ Retry-After headers - Proper 429 responses with retry guidance
• 💥 Fail-open - Guard outage doesn't block your production traffic
• 🔧 Auto-setup - Guard policy auto-creates on first request

🎯 How It Works

1. Webhook receives POST request

2. Identity extracted from headers:

• API key (x-api-key) → per-client limiting
• Client IP (X-Forwarded-For / x-real-ip) → per-IP limiting

1. Guard decides allow or deny:

• POST /api/v1/guard/{route:identity}/counter
• Checks against configured rate limit policy

1. Allowed → your business logic executes → 200 OK

2. Denied → immediate 429 Too Many Requests + Retry-After header

Client → Webhook → Identity → Guard → Allowed? → Business Logic → 200 OK
 ↓ NO
 429 + Retry-After

🔧 Setup Requirements

• Ainoflow - Sign up free for Guard API access. Free plan available.

That's it. One credential, one API.

⚡ Quick Start

1. Import workflow and set Ainoflow Bearer credential on GuardCheck node

2. Edit Config node with your limits:

Variable	Default	Description
rate_limit	30	Max requests per window
window_sec	60	Window in seconds
identity_mode	ip	ip or apiKey
route_name	webhook	Endpoint name

3. Replace BusinessLogic node with your workflow

Access original request:

const body = $('Webhook').first().json.body;
const headers = $('Webhook').first().json.headers;

4. Activate and test

🧪 Testing

Burst Test

Bash (Linux/macOS):

for i in {1..50}; do
 curl -s -o /dev/null -w "%{http_code}\n" \
 -X POST https://your-n8n.com/webhook/rate-limited-endpoint \
 -H "Content-Type: application/json" \
 -d '{"test": true}'
done

PowerShell (Windows):

1..50 | ForEach-Object {
 (Invoke-WebRequest -Uri "https://your-n8n.com/webhook/rate-limited-endpoint" -Method POST -Body '{"test":true}' -ContentType "application/json" -UseBasicParsing).StatusCode
}

Expected: First 30 → 200, remaining → 429

Proxy Test

curl -H "X-Forwarded-For: 1.2.3.4, 5.6.7.8" \
 -X POST https://your-n8n.com/webhook/rate-limited-endpoint

Identity key should use 1.2.3.4 (first IP from chain).

💬 Response Examples

Allowed (200 OK)

{
 "ok": true,
 "data": { "message": "Request processed successfully" }
}

Denied (429 Too Many Requests)

Headers: Retry-After: 17

{
 "ok": false,
 "error": "rate_limited",
 "retryAfter": 17
}

🏗️ Workflow Architecture

Section	Nodes	Description
Rate Limit Check	Webhook → Config → BuildIdentity → GuardCheck → IfAllowed	Extract identity, check Guard
Allowed Path	BusinessLogic → RespondOk	Your logic + 200 response
Denied Path	BuildDeniedResponse → RespondRateLimited	429 + Retry-After

Total: 9 nodes. Minimal by design.

🔒 What This Protects Against

• ✅ Webhook flooding - bot traffic, retry storms hitting your endpoint
• ✅ Credit burn - one runaway loop = €500+ OpenAI/Twilio bill overnight
• ✅ Automation overload - uncontrolled DB writes, external API hammering
• ✅ Accidental loops - webhook chains triggering each other endlessly

❌ What This Does NOT Replace

• Cloudflare / WAF (network-level protection)
• Bot detection (behavioral analysis)
• Layer 3/4 DDoS mitigation
• Authentication (who is the user?)

Guard handles application-level rate decisions, not network security.

🔑 Identity Modes

IP Mode (default)

Best for public webhooks where clients don't have API keys.

X-Forwarded-For: 1.2.3.4, 5.6.7.8 → identity = "1.2.3.4"
x-real-ip: 10.0.0.1 → identity = "10.0.0.1"

⚠️ IP addresses can be shared (NAT, mobile carriers, offices).

API Key Mode

Best for authenticated endpoints with per-client keys.

x-api-key: client_abc123 → identity = "client_abc123"

Falls back to IP if header is missing.

🛠️ Customization

Rate Limit Presets

Use Case	rate_limit	window_sec	Result
Burst protection	30	60	30/min
API rate limiting	100	3600	100/hour
LLM cost protection	10	60	10/min
Daily limit	1000	86400	1000/day

Multiple Endpoints

Use different route_name values to create separate rate limits:

Config A: route_name = "orders" → key = "orders:1.2.3.4"
Config B: route_name = "payments" → key = "payments:1.2.3.4"

Each route has independent counters.

Fail-Open vs Fail-Closed

Default: Fail-open - Guard API uses failOpen=true, so Guard outage doesn't block traffic.

To switch to fail-closed: change failOpen query parameter to false in GuardCheck node.

Combine with Shield (Dedup Protection)

Getting duplicate webhook deliveries? Add Ainoflow Shield before your business logic - one trigger, one execution, guaranteed. Guard + Shield = rate limiting + deduplication on the same endpoint.

⚠️ Important Notes

• Guard policy auto-creates on first request with rateMax/rateWindow parameters
• allowPolicyOverwrite=true is set for easy demo/testing - Config node values always apply. Production: set to false in GuardCheck query params to lock policy and prevent hidden config drift
• Denied requests not counted - only successful requests increment the counter
• Window resets atomically - no gradual decay, clean reset every N seconds
• No state in n8n - all rate limiting state lives in Guard API
• 5-second timeout - GuardCheck has 5s timeout to prevent blocking

💼 Need Customization?

Want to add temporary bans, cost protection mode, multi-tier rate limiting, or per-client usage dashboards?

Ainova Systems - We build custom AI automation infrastructure and safety layers for production workflows.

---

Tags: webhook, rate-limiting, security, guard, burst-protection, api-protection, ainoflow, production, webhook-security, cost-control