Right now the node-pcap parser just dumps the whole packet and headers into the netdata DB. This is kinda wasteful.
On startup, (Modeler?) should parse the installed model folders for, say, a manifest.json file that gives the required backends and fields from the model spec, merge them into a big tree, and pass that to Parser. (This will need updating the spec too.)
Then Parser slurps it up and includes every field it recognizes.
Right now the node-pcap parser just dumps the whole packet and headers into the netdata DB. This is kinda wasteful.
On startup, (Modeler?) should parse the installed model folders for, say, a
manifest.jsonfile that gives the required backends and fields from the model spec, merge them into a big tree, and pass that to Parser. (This will need updating the spec too.)Then Parser slurps it up and includes every field it recognizes.