Add more trackable metrics to understand why nodes are not connectable

[iberflow]

We should send details of NAT hole punching and port forwarding into monitoring.

• Metric on session failed with parameters on why it failed.
• Metric to measure how many nodes using Port forwarding.
• Don't forget to separate UPnP port forwarding from manual port forwarding.
• Extend metric of communication messaging with type information.