Timeouts handling

[rsvoboda]

#23 introduces new tooling which defines several timeouts

We need to think carefully about timeout and their handling. Not just timeouts for tooling, but also timeouts in tests.

For now I would stick with reasonable defaults, don't want to complicate this TS with infrastructure for dynamic timeouts at this moment.

We should be aware of this subject and have plans with possible solutions

[rsvoboda]

For remote connections following timeouts should be assessed

• connection timeout
• read timeout

For infrastructure we should be thinking about

• infrastructure images fetch timeout
• images boot timeout
• TS dependencies fetch time

Different environments -> different timeouts ?

• timeout multiplier
• retry approach

Infrastructure images fetch is probably the most time consuming item

• pre-fetch infrastructure images
• distribute infrastructure images (bundle of TS and all the required bits)