Aborted sessions don't kill requests to grid

[VitaliLiavitski]

If we aborte Jenkins job, requests have status in progress for using phones and simulators

[vdelendik]

it might be not a bug but let's investigate if we can handle somehow abort use-case on hub layer. By default aborted sessions should be disconnected by timeout.
Pay attention to the adb connection which might occupy device for longer time than timeout

[vdelendik]

fixed by #54.
testing scenario should be:

1. start iOS test or job via single device (ask @vdelendik to prepare explicit branch with rather slow steps)
2. kill iOS jobs in the middle and start timer to track delay
   -> iOS device is occupied either in mcloud/grid/console and in stf
3. start again iOS job vs the same device
   -> job is started, request added to queue as device still occupied in grid
4. tail appium device log to detect:

2022-01-24 23:02:25:519 - [HTTP] --> DELETE /wd/hub/session/60cd4cef-c063-4ecb-8beb-bf422598784c
2022-01-24 23:02:25:519 - [HTTP] {}
2022-01-24 23:02:25:521 - [debug] [W3C (60cd4cef)] Encountered internal error running command: NoSuchDriverError: A sessio                                                                                                                   n is either terminated or not started

-> as only previous session removed in appium (1-4 min) new iOS job should occupy device and start testing

So goal to identify that pause in occupation is upto 4 minutes instead of 60m as in v1.0

[vdelendik]

tested on my own