Adds proposal for Prometheus metrics

[aspic]

As mentioned in #46, adds initial support of Prometheus metrics. These metrics must then either be pushed, or as in this case be scraped from the endpoint /prometheus. This PR is merely to show what is possible, and only exposes a few of the metrics.

One must-fix if we decide to progress with this is to ensure /prometheus endpoint is "protected" somehow (at least if exposing anything like IP adresses).

A method of testing this can be to add this to the docker-compose-file (the proxy exposes metrics → prometheus scrapes metrics → grafana displays metrics):

nanorpcproxy:
....docker-compose-things

prometheus:
  image: prom/prometheus
  volumes:
    - /path/to/prometheus.yml:/etc/prometheus/prometheus.yml
  ports:
    - 9090:9090

grafana:
  image: grafana/grafana
  depends_on:
    - prometheus
  ports:
    - 3000:3000
  user: "472"

And make the prometheus.yml point to the Nano RPC proxy prometheus endpoint like this:

global:
  scrape_interval: 30s
  scrape_timeout: 10s
  
scrape_configs:
  - job_name: services
    metrics_path: /prometheus
    static_configs:
      - targets:
          - 'nanorpcproxy:9950'

When setting up Prometheus as datasource in Grafana, you would then get live metrics like this:

[Joohansson]

Sounds interesting! I guess you also have to add the extra setting to settings.json.default?
Did you leave that out by purpose?

[aspic]

Updated PR with these changes:

• Skips rate limiting of the /prometheus endpoint
• Prometheus is only enabled if the enable_prometheus_for_ips is set to avoid someone exposing the endpoint to the whole world. This address should then preferable only be on the internal network.
• Adds metrics for more stuff (rate limited, and external calls)