Open vdelendik opened 9 months ago
it has sense to test on production environments first this idea. Globally we wanna to achieve more stability even if any unique devices agents/nodes are corrupted. In this case sessions will be in queue until node is recovered, scaling env scaled or timeout for the job achieved... moving to 2.5 milestone as we'll adjust several prod envs to test on real use-cases first
In new approach with smart queue it has sense to keep requests even if required capability not present. One more benefit of this approach is that we can start thinking about dynamic devices approach 9externl hub providers, aws or emulators etc)
During the testing we have understand what's going on and what's exception is returned after
GRID_NEW_SESSION_WAIT_TIMEOUT=30000
it suppose to be error about node available....