Optionally expose usage metrics

[vruello]

Implements #189

This adds the following metrics exposed by a Prometheus endpoint: https://github.com/vruello/openwec/blob/metrics-rs/doc/monitoring.md#available-metrics

Machine labels can be added for any supported metric in the monitoring settings:

# [monitoring]

# [Required]
# Listen address of the Prometheus-compatible endpoint
# listen_address =

# [Required]
# Listen port of the Prometheus-compatible endpoint
# listen_port =

# [Optional]
# The refresh interval of "openwec_machines" gauge
# machines_refresh_interval = 30

# [Optional]
# Request duration buckets (in seconds) used by the "openwec_http_request_duration_seconds" histogram
# http_request_duration_buckets = [0.005, 0.01, 0.025, 0.05, 0.1, 0.25, 0.5, 1.0, 2.5, 5.0, 10.0]

# [Optional]
# If set, a "machine" label will be added to the "openwec_input_events_total" metric
# Warning: this may cause a HUGE increase in metric cardinality
# count_input_events_per_machine = false

# [Optional]
# If set, a "machine" label will be added to the "openwec_input_event_bytes_total" metric
# Warning: this may cause a HUGE increase in metric cardinality
# count_input_event_bytes_per_machine = false

# [Optional]
# If set, a "machine" label will be added to the "openwec_http_request_body_network_size_bytes_total" metric
# Warning: this may cause a HUGE increase in metric cardinality
# count_http_request_body_network_size_per_machine = false

# [Optional]
# If set, a "machine" label will be added to the "openwec_http_request_body_real_size_bytes_total" metric
# Warning: this may cause a HUGE increase in metric cardinality
# count_http_request_body_real_size_per_machine = false

[vruello]

This seems stable enough to be merged soon