Instrument goProbe with metrics

[els0r]

For #74 and related issues (especially the performance relevant once), it has become increasingly clear that we need more eyes. This issue follows up on #160 to make sure we collect more performance relevant telemetry from goProbe.

The will be of the form goprobe_<sub system>_<name>

Name	Type	Description
Capture
goprobe_capture_packets_processed_total	Counter	Counts the total number of packets processed across all interfaces. Modified every 5 minutes as part of the stats collection during writeout
goprobe_capture_packets_dropped_total	Counter	Counts the total number of packets dropped across all interfaces. Modified every 5 minutes as part of the stats collection during writeout
goprobe_capture_errors_total	Counter	Counts the total number of capture errors across all interfaces. Modified on demand when error occurs
Capture Manager
goprobe_capture_manager_writeout_duration_seconds	Histogram	Observes the total writeout time. Buckets: 1ms, 5ms, 10ms, 25ms, 50ms, 100ms, 250ms, 500ms, 1s, 5s
goprobe_capture_manager_interfaces_capturing_total	Gauge	Counts the number of interfaces that are currently capturing traffic
API
goprobe_api_request_duration_seconds	HistogramVec	Observes the request duration. Buckets: 10ms, 50ms, 100ms, 250ms, 1s, 5s, 10s, 30s, 60s, 300s. The lower range is for status commands, the upper ones for invocations of _query. Will be refined for global-query in separate issue. Labels: host, method, path, status_code

DoD:

• set up /metrics endpoint in goProbe
• implement the main metrics and register them

Bonus:

• Investigate the use of vectors in case interface cardinality is low (make optional)