OK,Dumb Q, but what's the advantage of one over the other between using a) the built in monitoring

service:
  http_server: on
  http_listen: 0.0.0.0
  http_port: 2020

https://docs.fluentbit.io/manual/administration/monitoring vs https://docs.fluentbit.io/manual/data-pipeline/inputs/fluentbit-metrics and sending those to a prometheus_exporter as described in the docs? Does one have more things? better coverage of pipeline/ consume fewer resources?

The http server opens more stuff up potentially plus you can't have mixed Auth, e.g. no Auth on health but Auth on other endpoints. That's the way I've looked at it: if you just need metrics then keep it minimal

Ta. I was also looking to see what dashboards are already out there for picking up slowdowns / trends, and that's when I realised that https://github.com/fluent/fluent-bit-docs/blob/master/monitoring/dashboard.json wasn't really working with just the metrics (also we use victoria metrics native datasource not the prometheus variant, so needed to do a quick search/replace on the json before it'd work at all.