refactor: merge metrics with PromEx

[avilagaston9]

Motivation

We were using two Prometheus clients (prometheus_ex and telemetry_metrics_prometheus).

Description

Merge our BEAM and custom metrics using PromEx:

• Configures the PromEx module to start a standalone HTTP server to serve the aggregated metrics .
• Removes prometheus_ex, telemetry_metrics_prometheus and their dependencies.
• Sets the PromEx.Beam plugin, which captures metrics regarding the Erlang Virtual Machine.
• Exports the beam.js pre-built dashboard and uses it.
• Defines a custom plugin with all the metrics that we were using for the node dashboard.
• Removes the old BEAM dashboards since they are replaced by the pre-built one.
• Modifies the --metrics and --metrics-port flags to have the same logic as the beacon-api flags.

As a result, it is no longer necessary to start the BeaconApi to see the BEAM metrics. They are all aggregated in the PromEx server.

Closes #903

[rodrigo-o]

LGTM!

A small comment for kurtosis, given that we need to setup the dashboards with make kurtosis.setup.grafana for it to be taken, the current version in main of the make task just add the new dashboards without removing the old ones, we could make this change to take that into account:

kurtosis.setup.grafana:
	[ -d  $(KURTOSIS_GRAFANA_DASHBOARDS_DIR)/lambdaconsensus ] && \
		rm $(KURTOSIS_GRAFANA_DASHBOARDS_DIR)/lambdaconsensus/* || \
		mkdir $(KURTOSIS_GRAFANA_DASHBOARDS_DIR)/lambdaconsensus
	cp -r $(GRAFANA_DASHBOARDS_DIR)/* $(KURTOSIS_GRAFANA_DASHBOARDS_DIR)/lambdaconsensus