Technical Specifications
The numbers behind the service — limits, latency, and guarantees.
Message Limits
- Max size: 4KB / 16KB / 64KB / 256KB by plan
- Payload format: JSON only
- History: 1 / 3 / 7 / 14 days by plan
Messages exceeding the size limit will be rejected.
Rate Limits
- Publish: 20 / 100 / 300 / 1,000 msgs/sec by plan
- History reads: 1,000 requests/min
- Excess publishes return HTTP 429
Per-second publish rates are enforced in real time. Monthly message quotas are soft — exceeding them triggers notifications, never a mid-session cutoff.
Latency
- p50: <50ms
- p95: <150ms
- p99: <300ms
Measured from publish API call to client message receipt.
Message Ordering
- Per-channel FIFO ordering guaranteed
Messages within a single channel are delivered in the order they were published.
Delivery Semantics
- At-most-once delivery
Messages are delivered once or not at all. Use message history for guaranteed delivery.
Infrastructure
- 300+ Cloudflare cities
- Automatic failover
- 99.9% uptime SLA (Starter tier)
Deployed globally on Cloudflare's edge network.
Reconnection
- Automatic reconnection
- Exponential backoff (1s initial, 30s max)
- Message replay on reconnect
The client SDK handles disconnections automatically.
