Dynamic inventory: Error handling and debugging must be improved

[jbuchhammer]

Description

As it showed up in issues #26 and #27 error handling in profitbricks_inventory.py is rudimentary. Only a few parts are covered by try/except clauses and debugging information on the context of the error is missing completely. As a user I have to manually edit the inventory script to find the ressource where the error occurred and get more details.

Expected Behaviour

An error handling must be implemented that

• provides detailed information on the context of the error (e.g. affected resource)
• catches errors early
• automatically tries to recover from errors if possible, e.g. wait and retry if API calls return a PBRateLimitExceededError or skip resources that are not suitable (as in Dynamic inventory fails if a server has no boot device #27).

In addition, a logging with configurable log level would be helpful.