I've been banging my head against a brick wall for the better part of 2 days trying to get this godforsaken node graph to correctly display data. The examples given by Grafana are essentially useless, since I can't think of a single instance when I would just want static CSV data in a node graph. I want dynamic data! Well there's virtually zero documentation on how to actually achieve this, despite finding many posts of people asking the same questions. My confusion is this. t * Nodes and Edges support mainstat and secondarystat * But a prometheus query can only return one metric at a time * Using one query to grab mainstat and another query to grab secondarystat means you lose the singular "nodes" query necessary to fill out the graph * I can use transformations to UNION these queries into one dataframe, but this does not end up as "nodes" but some other refId If I try and simplify and only use a mainstat, I run into another issue. Prometheus returns the default "Value" for the metric, but no column named "mainstat". And the \*exact\* transformation I would need to create that column (Organize Fields By Name) is conveniently greyed out. It works on the UNIONed table, but again, It's no longer called "nodes" so no longer appears on the graph. It seems like a spiderweb of catch 22s where I can't nail down a query/transformation that actually gives me what I want. Here's what I have so far. A query to generate the "nodes" dataframe group by (id, title) ( label_replace( label_replace( ( # Include traefik as a node label_replace(vector(1), "service", "traefik", "", "") or # Include all services that traefik routes to group by (service) (traefik_service_requests_total) ), "id", "$1", "service", "(.*)" ), "title", "$1", "service", "(.*)" ) ) This outputs data in the form `{"", id="grafana@file",title="grafana@file"}` Then I have my edges query group by (id, source, target) ( label_replace( label_replace( label_replace( sum by (service) (rate(traefik_service_requests_total[$__range])), "source", "traefik", "", "" ), "target", "$1", "service", "(.*)" ), "id", "traefik-to-$1", "service", "(.*)" ) ) Which produces rows like `{id="traefik-to-grafana@file", source="traefik", target="grafana@file"}` This does successfully draw nodes and lines. But there are no actual values associated with these queries, this just gives me the nodes and edges.