OK, I don’t see any difference between your config and mine. Other than my Prometheus/Grafana server is a 4th machine. I changed Prometheus IP to 0.0.0.0, but that doesn’t change the fact that it is only exporting 19 “cardano.node.*” metrics on port 12798. I did find a report of this issue, and it said to map the old namespaces, but that doesn’t solve my issue either.
One other thing we tried: we found a note that turning on some tracers outputs some stats to the Prometheus output (see Sdoutriaux reply towards the end of the thread), so we turned on ALL of the tracers that were turned off and restarted the node. That only added 2 more stats, but none of the Forge or ChainDB ones. Maybe the stat names are changed now?
I think I am onto something. I noticed for example that one dashboard panel reports on cardano_node_ChainDB_metrics_slotNum_int, which doesn’t exist, but I did find cardano_node_metrics_slotNum_int. So I think the change in namespace must not have been reflected from this IOHK dashboard I downloaded. However, the same is not true for these metrics:
cardano_node_metrics_Forge_node_is_leader_int - I don’t see any “leader” metrics exported.
There are other panels that use *Forge* metrics, but I can’t find them by the last elements in the namespace (e.g. “node_is_leader”). So half of my issue is resolved by the fact that the ChainDB metrics were renamed.