For the recent week, the team have been facing the same problem about deploying pods, since the team did not choose to reserve GPUs so it is understandable that we were often shown "unavailable" under most GPUs.
However, when the time is right and we finally get to deploy a pod using lets say RTX 4090, the pod is successfully deployed with a little green dot showing it's active, it is actually un-connectable, the browser page would went error again and again, few successful deployment where we finally connect to it takes more than 6 hours of reloading the browser page, we have tried different browsers like chrome and firefox and it doesn't help at all.
According to the top right counter the team have been paying for all the waiting time. which is obviously a waste.
questions
1. are we alone in this? is it our connection problem or somebody else's?
2. do we fix it if we switch to multi-month saving plans that reserve a GPU instead of on-demand deployment?