Accumulated timeouts handling

[nieznanysprawiciel]

Description:

• If external layer has timeout on message response recival, we might not have enough time to initialize session.
• Each operation (querying node information, establishing session, tcp connection) adds additional timeout, which can accumulate (each operation succeeds, but closer to upper margin of timeout)
• Previous implementation was trying all possible methods to establish session. If direct p2p to other Node failed, Reverse Connection was attempted and relayed connection was a fallback. This accumulates timeouts even more
• First yagna call to other Node can timeout despite session will be initialized later correctly (in most cases it will be counter proposal, but very often sending DebitNote/Invoice will be done after longer period of waiting, when session needs to be re-established)

What:

• Maybe there exist solution to this problems
• At least re-evaluate all timeouts to find optimal values
• Consider which methods to establish session we should try after each other and when we should just give up with error

[staszek-krotki]

Research needed.