Install otel-collector to collect broker and server metrics

[tbroden84]

Make it optional to install otel-collector with configurable endpoints.
Additional tags to tag each run by name and LavinMQ version.

Reference: #24

[jage]

We probably want a pretty high scrape rate here, like every 15s?

[jage]

Would be nice with peer latency (i.e. idle network latency between generator and broker)

[jage]

One idea is to not send to Grafana. Now that we have graph visualisations maybe we should continue with CSV files, that's also easier to reference when comparing really old runs etc.

Lets install otel-collector, but we build a little tool that scrapes /metric both for broker metrics and host metrics. We just store the raw values as CSV files with timestamps, preferably make sure our benchmark results also have timestamps if they run over time.

#33 would be a good scenario to implement together with this, how much RAM does each connection use.