nautechsystems/nautilus_trader · error
No healthy transport clients available
Error message
No healthy transport clients available
What it means
broadcast_cancel collects the transport clients flagged healthy and aborts if none remain, incrementing the failed_cancels metric. This means every BitMEX transport connection is currently unhealthy (disconnected, reconnecting, or failed health checks), so there is no channel over which to send the cancel request.
Solutions
- Check/restore network connectivity and wait for transport reconnect before retrying the cancel
- Verify transports are connected and authenticated before submitting/cancelling orders (gate on a connected event)
- Inspect health-check and reconnect logic/logs to see why all clients were marked unhealthy
- Fall back to the REST cancel endpoint if available
- If cancels can fire at startup, add a readiness check or startup delay
Example fix
// before
let result = canceller.broadcast_cancel(&cl_ord_id).await?;
// after
if !transport.is_healthy() {
transport.wait_until_healthy(timeout).await?; // or fallback to REST cancel
}
let result = canceller.broadcast_cancel(&cl_ord_id).await; Defensive patterns
Strategy: retry
Validate before calling
// guard before cancelling
if !transports.iter().any(|t| t.is_healthy()) {
return Err(NoHealthyTransport); // or wait/reconnect first
} Type guard
fn has_healthy_transport(transports: &[Transport]) -> bool {
transports.iter().any(|t| t.is_healthy())
} Try / catch
match canceller.broadcast_cancel(&cl_ord_id).await {
Err(e) if e.to_string().contains("No healthy transport clients") => {
wait_for_reconnect(timeout).await?;
rest_client.cancel_order(&cl_ord_id).await?; // fallback
}
other => other?,
} Prevention
- Gate order operations on a connected/ready signal
- Implement REST fallback for cancels when websockets are down
- Monitor transport health metrics and alert on total unhealth
- Verify authentication succeeds at startup so clients become healthy
When it happens
Trigger: Calling broadcast_cancel(order_id) while all transports report is_healthy() == false — e.g. before the first connection is established, during an outage, or after all clients hit the failure threshold and were marked unhealthy.
Common situations: Network drop or BitMEX outage at the moment a strategy issues a cancel; startup ordering where cancels fire before transports connect; websocket auth failure marking all clients unhealthy; firewalled/restricted network blocking wss://www.bitmex.com.
Related errors
- Asks length mismatch: expected
- Bids length mismatch: expected
- Both `api_key` and `api_secret` must be provided together
- Missing is_inverse for
- Missing is_quanto for
AI-assisted analysis of nautechsystems/nautilus_trader@18893faf8b (2026-09-08).
Data as JSON: /api/errors/0f55f7e48de30a0a.
Report an issue: GitHub.
Appendix: source
Thrown at crates/adapters/bitmex/src/broadcast/canceller.rs:639
/// Returns an error if all cancel requests fail or no healthy clients are available.
pub async fn broadcast_cancel(
&self,
instrument_id: InstrumentId,
client_order_id: Option<ClientOrderId>,
venue_order_id: Option<VenueOrderId>,
) -> anyhow::Result<Option<OrderStatusReport>> {
self.total_cancels.fetch_add(1, Ordering::Relaxed);
let healthy_transports: Vec<TransportClient> = self
.transports
.iter()
.filter(|t| t.is_healthy())
.cloned()
.collect();
if healthy_transports.is_empty() {
self.failed_cancels.fetch_add(1, Ordering::Relaxed);
anyhow::bail!("No healthy transport clients available");
}
let mut handles = Vec::new();
for transport in healthy_transports {
let handle = get_runtime().spawn(async move {
let client_id = transport.client_id.clone();
let result = transport
.cancel_order(instrument_id, client_order_id, venue_order_id)
.await
.map(Some); // Wrap success in Some for Option<OrderStatusReport>
(client_id, result)
});
handles.push(handle);
}
self.process_cancel_results(
handles,View on GitHub (pinned to 18893faf8b)