Exceeding carrier TPS caps causes aggregator queue latency or instant message rejections. Implement token bucket rate limiters to pace outbound requests cleanly.
Telephony connections are bounded by rigid throughput limits measured in Transactions Per Second (TPS). Standard 10DLC numbers are restricted to 10-30 TPS per carrier, while toll-free lines allow 3-50 TPS. Disregarding these ceilings causes messages to back up in aggregator queues, ballooning OTP delivery latencies.
1 Token Bucket Rate Limiting Architecture
Implement a distributed Token Bucket rate limiter (e.g. using Redis or Cloudflare Workers KV) at the application dispatch layer. Ensure outbound API calls never exceed 80% of your provisioned carrier TPS quota to preserve headroom for latency spikes.
Frequently Asked Questions • Carrier Spam Filters & Deliverability
Aggregators will buffer the excess 90 messages per second. After 30 seconds of sustained overflow, queue latency will exceed 30 seconds, causing OTP timeouts.