Can Produce guarantee that events are processed and delivered in order even under failure?

[artog]

My team is building an application in C# that reads from one Kafka cluster and writes to another. It is vital that:

1. Events are written in order
2. No event is skipped

Initially we used an awaited ProduceAsync call but were concerned about speed. It goes much faster when using Produce, but we have noticed that the code continues after calling it, before the delivery report has been received.

Does this mean that we continue producing entries even though we don't know the status of earlier entries?
Can we stop producing entries when some event fails?

Initially we thought about committing to to the source Kafka in the delivery report, but we are concerned about committing events after one has failed or otherwise had an error. Is there any guarantee that the delivery report will:

1. always be made?
2. always be in order?

What happens if the report is unable to be received, is the entry still written into the Kafka?