1. Do we have stats on liquidity in the RI marketplace (paraphrasing). Yes we do! We know that some instance types are more liquid than others and factor that into the Autopilot algorithm.
2. Does 50% account for risks? It does. We profile peaks and valleys in usage on a per compute class basis and always ensure that our RI coverage is optimally covering those peaks and valleys without running the risk of over-commitment. If there is a significant delta between peak and valley, we may make a lower RI commitment for that compute class to decrease risk. The specific coverage and savings rates are ultimately dependent on the usage by compute class which is different for each customer.
3. SOC2 timeline? We're expecting this before the end of the year.
4. Take into account existing RIs and SPs? Yes we do. The Autopilot dashboard will show you how much coverage you already have. And then here's a specific docs link with more details on this https://docs.vantage.sh/autopilot/#i-have-existing-aws-savin...