Fix connectivity between prometheus and prometheus-node-exporter

Expected to have regressed after https://github.com/spantaleev/matrix-docker-ansible-deploy/pull/1008

This patch comes with its own downsides (as described in the comments
for matrix_prometheus_node_exporter_container_http_host_bind_port),
but at least there's:
- no security issue
- metrics remain readable from matrix-prometheus (even if the network metrics are inaccurate)

A better patch is certainly welcome.

roles/matrix-prometheus-node-exporter/defaults/main.yml

@@ -19,4 +19,16 @@ matrix_prometheus_node_exporter_systemd_wanted_services_list: []
 # Controls whether the matrix-prometheus container exposes its HTTP port (tcp/9100 in the container).
 #
 # Takes an "<ip>:<port>" value (e.g. "127.0.0.1:9100"), or empty string to not expose.
+#
+# Official recommendations are to run this container with `--net=host`,
+# but we don't do that, since it:
+# - likely exposes the metrics web server way too publicly (before applying https://github.com/spantaleev/matrix-docker-ansible-deploy/pull/1008)
+# - or listens on a loopback interface only (--net=host and 127.0.0.1:9100), which is not reachable from another container (like `matrix-prometheus`)
+#
+# Using `--net=host` and binding to Docker's `matrix` bridge network may be a solution to both,
+# but that's trickier to accomplish and won't necessarily work (hasn't been tested).
+#
+# Not using `--net=host` means that our network statistic reports are likely broken (inaccurate),
+# because node-exporter can't see all interfaces, etc.
+# For now, we'll live with that, until someone develops a better solution.
 matrix_prometheus_node_exporter_container_http_host_bind_port: ''