Option for only exporting varnish metrics

[nipuntalukdar]

Added a command line parameter for -no-go-metrics to switch on or off exporting go runtime and http handler metrics.

[nipuntalukdar]

Added changes for adding varnish environment and ip address as labels to the metrics. This will be useful when we collect metrics from multiple varnish servers.

[jonnenauha]

Was going to merge before the last commit. But now this is getting a bit messy.

What is the use case for env labels from command line? IP address is automatically available to you from prometheus and you can add labels to the scrape config like "production" "staging". That is a feature of prometheus already, attaching labels that will be useful to you when creating queries to certain servers. I don't see the benefit of doing it in the exporter? It seems the code adds the labels into all exported mertics, that is exactly what the prometheus yml config is for :)

Here is a example. IP address is already in the standard instance.

  - targets: [ '<ip1>:9131', '<ip2>:9131' ]
    labels:
      service: 'varnish'
      frontend_for: 'proxy'
      whatever_data_you_need: 'dev'
      kind: 'stable' // or production or whatever you need

If you have multiple servers, you make multiple entries to the prom config with different labels. Its a lot convenient than going to the server and configuring them there.

[nipuntalukdar]

Will update and create a new pull requests. I didn't know about the options for automatically adding the labels for ip etc.:) as I am new to Prometheus.