Feature: health check for peers

[bitromortac]

Introduce command peer-check [nodeids]:

• if we are not connected, we should try to connect, the peer should reply in a reasonable time
• make active part asynchronous
• they should quickly respond to network pings (implement also onion pinging)
• pings from the listpeers command should be taken into account
• they should have a good uptime
• they should send back funds on a minimal circular rebalance
• error messages and flap counts need to be taken into account
• try to trigger an error message and measure response time
• incorporate number of routing failures from previous rebalancing attempts
• relative badness
• response time from rebalancing
• peers in common with own peers
• incorporate last updated from channel / node announcements
• how many liquid channels does the peer have
• how often could the peer route a payment relatively speeking
• how many updates were done to the channel(s)
• how many channels does the peer have in common with our peers