Configuration of the ingester is definitely a sore spot. Some ideas:
the "schema" field in the message can be used to suggest a plugin, which could then be fetched/loaded and used to archive data. Because of the protobuf-json translation, we could embed jq and have parsing logic defined in terms of jq programs. We aren't currently using the schema field for anything other than xbosproto/XBOS. Could also put this in the XBOS top level message
the "contextd" daemon saves an RDF representation of published resources. This engine could be integrated into the ingester and used to bootstrap the subscriptions
add default subscriptions to the ingester.yml file
Configuration of the ingester is definitely a sore spot. Some ideas:
xbosproto/XBOS. Could also put this in the XBOS top level messageingester.ymlfile