[maybe] Return workflowStatus in the context so that RPC can avoid using write API

This will make it easier for implementing RPC, to avoid having to catch the "WorkflowNotExsits" error and call another RPC.

However, when status is running, this cannot guarantee it is still running when the worker is executing the RPC......So may not worth doing this.

Start by tracing the RPC entry point and how workflow status is currently obtained, including the WorkflowNotExsits path and the write API. Determine whether returning workflowStatus in the context can avoid that extra RPC without creating a stale-status assumption. Done means the approach and its behavior for running and missing workflows are clearly defined.

Written by the indexing model from the issue text.