status-wda check should be as fast as possible to avoid hub down

zebrunner / mcloud-grid

Customized selenium-hub

6 stars 3 forks source link

status-wda check should be as fast as possible to avoid hub down #72

Closed vdelendik closed 2 years ago

vdelendik commented 2 years ago

it was discovered that every status-wda call might wait upto 60s (default http client timeout). it means if we have 3-4 crashed wda we spent 3-4 extra minutes to find node... it should be fixed asap

vdelendik commented 2 years ago

merged: https://github.com/zebrunner/mcloud-grid/pull/71

vdelendik commented 2 years ago

verified on my own killing wda process forcibly. we have double shield now: stf and device state itself and extra status-wda.