How can I obtain notifications of bucket replication failure events?

[feixiangyimeng]

Whether using a webhook or MQTT subscription, the obtained REPLICA status is always pending, and the failed status is never obtained. Setting the replication bucket to read-only also does not yield a failed status. Unless I write a timed polling function to determine failure if the replication bucket does not retrieve the object within a timeout, is there any other way?

无论是webhook还是MQTT订阅，获取的REPLICA 状态一直是pending，得不到failed状态，把复制桶设置为只读也得不到failed,除非自己写个定时轮询在复制桶如果超时没有获取到对象则判定失败，有没有其他方法？

[For8stt]

Afaik there is no separate “replication failed” event emitted right now, or at least it does not look like Webhook/MQTT exposes it reliably.

If the event only stays in pending, then relying on notifications alone is probably not enough for failure detection.

For now I’d handle this with polling / timeout logic, e.g.:

1. receive replication pending
2. check destination bucket/object after N seconds/minutes
3. if object is still missing or metadata does not match, mark it as failed on the app side

Also worth checking server logs / metrics, because the actual replication error may be visible there even if the event status is not updated to failed.

Imo either:

• failed events are not implemented/exposed yet, or
• this is a bug in the replication notification status update

Would be good to know if RustFS plans to emit explicit replication:failed / REPLICA failed events for Webhook/MQTT. Right now timeout-based detection seems like the safest workaround.