Kafka functions: KEDA scaling with consumer-lag trigger

`func deploy` already supports `--deployer keda` for KEDA-based scaling. For Kafka functions, add a Kafka consumer-lag trigger so KEDA can scale based on consumer group lag — enabling scale-to-zero when there are no messages and scaling up under load. Currently Kafka functions require `min-scale: 1` to stay running.

Traffic-based autoscaling (e.g. Knative's KPA or KEDA HTTP scaler) does not work for Kafka functions because there is no inbound HTTP traffic — the function pulls messages from Kafka internally. Scaling must be driven by Kafka consumer group lag instead.

Start at the func deploy entry point for --deployer keda and trace how Kafka functions are represented in the deployment configuration. Compare the existing KEDA path with the Kafka consumer-group settings; done means KEDA can use consumer-group lag to scale Kafka functions to zero and up under load.

Written by the indexing model from the issue text.