nautechsystems/nautilus_trader · critical
Error reading from stream
Error message
Error reading from stream: {e:?} What it means
stream_messages reads entries from a Redis stream in a loop. If XREAD fails with a non-retryable error (is_retryable_stream_error returns false), the loop aborts by propagating this error with the debug representation of the underlying redis error. Retryable errors instead trigger reconnection logic and logging.
Solutions
- Inspect the wrapped error ({e:?}) to identify the exact Redis error (NOGROUP, WRONGTYPE, auth, etc.)
- If NOGROUP: recreate the consumer group (XGROUP CREATE ... MKSTREAM) and restart the consumer
- If WRONGTYPE: check what is writing to the stream key — delete/rename the conflicting key and fix the writer
- If auth/ACL errors: update the connection credentials or grant the user stream read permissions
- Restart the message bus after fixing the server-side condition; the loop does not self-heal non-retryable errors
Defensive patterns
Strategy: retry
Try / catch
match bus.stream_messages().await {
Err(e) if e.to_string().contains("Error reading from stream") => {
// inspect debug payload; recreate consumer group / fix key then restart with backoff
tokio::time::sleep(Duration::from_secs(5)).await;
}
other => other?,
} Prevention
- Ensure the consumer group exists (XGROUP CREATE ... MKSTREAM) before starting the bus
- Never write a different type to the stream key; reserve the key for the bus
- Use stable credentials/ACLs that grant stream read for the bus user
- Monitor for stream key deletion (FLUSHALL/DEL) in shared Redis instances
When it happens
Trigger: A persistent failure while calling XREAD on the stream — e.g. NOGROUP (stream deleted/consumer group missing), wrong-type errors (key holds a non-stream value), auth failures after connect, or protocol corruption that the retry classifier deems unrecoverable.
Common situations: The stream key was deleted or FLUSHALL'd while the bus was running; another client overwrote the key with a different type; Redis restarted with the consumer group gone; ACL/permission changes mid-session.
Understand the failure class
Background: Database query failed: Internal Server Error 500s wrapping SQL, Prisma, and connection failures — what to check first — this error's family across 16 libraries.
Related errors
- All requests failed
- All requests failed
- All event slug requests failed
- All slug requests failed
- Ambiguous order-list submit failure, awaiting reconciliation
AI-assisted analysis of nautechsystems/nautilus_trader@18893faf8b (2026-09-08).
Data as JSON: /api/errors/1c18d7b159abd538.
Report an issue: GitHub.
Appendix: source
Thrown at crates/infrastructure/src/redis/msgbus.rs:702
match decode_bus_message(array) {
Ok(msg) => {
if let Err(e) = tx.send(msg).await {
log::debug!("Channel closed: {e:?}");
break 'outer; // End streaming
}
}
Err(e) => {
log::error!("{e:?}");
}
}
}
}
}
}
}
Err(e) => {
if !is_retryable_stream_error(&e) {
anyhow::bail!("Error reading from stream: {e:?}");
}
log::error!("Error reading from stream: {e:?}");
let Some(reconnected) =
reconnect_stream_connection(&config, &stream_signal, &mut read_error_count)
.await?
else {
break;
};
con = reconnected;
}
}
}
log_task_stopped(MSGBUS_STREAM);
Ok(())
}View on GitHub (pinned to 18893faf8b)