We have a Druid cluster ingesting some 40k messages per second over about 50 different Kafka topics. The definition of which data to extract is different for these topics and extraction for individual topics change regularly, which means we need to restart our tranquility nodes to have this change take effect, risking all ingestion. Is there any way to verify the configuration before restart (e.g. a pre-flight check a la apachectl/nginx -t)? If not (the source code indicates not) would such a pull request be welcome?