You have some code for your Actor, not yet built nor deployed. You get no insights regarding gas usage.
You deploy your code to the network, regardless of whether it is local devnet or mainnet. Though we may enforce putting it only on devnet if we cannot get enough insights from mainnet.
Once the Actor is deployed and some calls are made and traced, we can mark which method uses how much gas, what's the % overall across "tracked" actors, and how many times it was called and visualise it, either in a separate floating window, directly under the code or in some other ways. The LSP is fed information through the API Gateway/Monitoring, so this should be pluggable with another tooling that someone may devise.
How I would see it, naively, is as follows: