2 個答案
- 最新
- 最多得票
- 最多評論
1
In a distributed system, race conditions can occur that make this possible, or better said, it is not impossible. Idempotency is the solution.
已回答 1 年前
0
SQS was designed with the architecture you describe in mind - auto-scaled fleets of worker instances all consuming the same queue are a very common pattern. Receiving the same message is rare (I've never seen it when looking through logs); there's probably no guarantee though and this possibility is covered by the declared at-least-once delivery.
相關內容
- AWS 官方已更新 1 年前
- AWS 官方已更新 7 個月前
- AWS 官方已更新 2 年前
With standard queues this can happen. Maybe not a lot, but sure can.