Useful thanks. Some questions - 1: Why would you choose to use an eventstream over direct ingestion into KQL? Is it purely for transformation purposes? 2: Does the advent of eventhouses change your approach in certain scenraios in any way?
Thanks for feedback. To answer your questions: 1. Eventstreams could be helpful if your source isn't one that can feed to KQL directly (say IOT hubs) or if your transformations are simple enough to be done from it, otherwise it'll only add up to latency 2. Eventhouses are nice feature designed to optimize costs and make data sharing easier. However, at the base of Eventhouses are KQL databases, so all ingestion/transformation related to KQL are true for Eventhouses as well.
Useful thanks. Some questions -
1: Why would you choose to use an eventstream over direct ingestion into KQL? Is it purely for transformation purposes?
2: Does the advent of eventhouses change your approach in certain scenraios in any way?
Thanks for feedback. To answer your questions:
1. Eventstreams could be helpful if your source isn't one that can feed to KQL directly (say IOT hubs) or if your transformations are simple enough to be done from it, otherwise it'll only add up to latency
2. Eventhouses are nice feature designed to optimize costs and make data sharing easier. However, at the base of Eventhouses are KQL databases, so all ingestion/transformation related to KQL are true for Eventhouses as well.