Are there overheads for Project Lucy and what are the benefits of having it activated ?
The data that is uploaded to Project Lucy is data that is already collected by the Diagnostic Server, so there is no additional overhead in turning Project Lucy upload on. Though Project Lucy can accept SQL Server Trace data as an input, Spotlight does not initiate a trace for the purposes of uploading it to Project Lucy. If the SQL Analysis feature is enabled in Spotlight as part of normal monitoring, that data will be uploaded, but again there is no additional overhead in taking advantage of the Project Lucy feature – Spotlight just uploads data it is already collecting.
The analysis done does try to work out whether the sample comes from a period of high load vs. periods of low load (or in fact different load). There are various statistical techniques used to work this out and these will surface in Project Lucy over time as the service matures.
When signing into Project Lucy, there will be a list of SQL Server Instances (or Oracle Databases) that belong to the end user that logs in. On each of these, a performance health for today is calculated. The general inanition is that the user will see all of his or her own data with comparisons to others where it makes sense to do so.