Nice post, i guess scaling the shards should also be based on the processing capacity of your…
yaron haviv

Good point, it *could* be, but a much harder judgement call to make as you can have many Lambda functions all processing events at different speeds.

Luckily for us, none of our Lambda functions are doing anything that are slow enough for that to be a consideration. In which case, auto-scaling on the ingress side was sufficient.

Show your support

Clapping shows how much you appreciated Yan Cui’s story.