Rate Limits

Mist enforces the following limits per API token:

Exceeding either limit can result in 429 responses during connect/subscribe flows and may trigger connection drops.

## Distribute Channels Across Multiple Connections

Never subscribe to thousands of channels on a single WebSocket. A single connection with too many channels means:

• A single point of failure for all your data
• Message backlog builds faster than it can drain
• Higher risk of hitting the 2,000 channel limit

❌ Bad — 2000 sites, 1 WebSocket, 2000 channels:

ws://api-ws.mist.com  →  /sites/site_1/stats/devices
                      →  /sites/site_2/stats/devices
                      →  ... (2000 channels)

✅ Good — 2000 sites, 30 WebSockets, ~70 channels each:

ws_1  →  /sites/site_1/stats/devices ... /sites/site_70/stats/devices
ws_2  →  /sites/site_71/stats/devices ... /sites/site_140/stats/devices
...
ws_30 →  /sites/site_1931/stats/devices ... /sites/site_2000/stats/devices

This also gives you natural parallelism and resilience — a single dropped connection only affects a subset of your data.

Handle 429 ResponsesCopy Anchor Link

Your reconnect logic must handle rate limit responses gracefully. If you receive a 429, back off before reconnecting — reconnecting immediately will just hit the limit again:

RECONNECT_DELAY = 5  # seconds, increase if hitting 429 frequently
def on_close(ws, code, msg):
    logger.info(f"WebSocket closed — code: {code}, msg: {msg}")
if __name__ == '__main__':
    while True:
        try:
            establish_ws()
        except Exception as e:
            logger.error(f"Connection error: {e}")
        time.sleep(RECONNECT_DELAY)

Plan for Message Volume, Not Just Channel CountCopy Anchor Link

The 2,000 channel limit is one constraint, but message volume per channel matters equally. A single site with thousands of clients can generate hundreds of messages per second on its own. If too many high-traffic channels are on one connection, the receive buffer fills up and starves the ping/pong mechanism — even if you're well under the channel limit.

Before finalizing your subscription layout:

• Estimate the total message rate across all channels on a single connection
• If you expect high traffic sites or large client counts, distribute aggressively — fewer channels per connection
• Monitor queue depth at runtime (see Message Throttling & Queue Monitoring) and rebalance if it grows unbounded

Need Higher Limits?Copy Anchor Link

If your use case legitimately requires more than 2,000 connections/hour or 2,000 channels/connection, contact Mist support to request a limit increase.