Side-effect-free way of checking on task status

Per a conversation on slack, there is no current side-effect-free way of checking on task status. This makes it difficult as an end user to write correct, reliable code, especially when `get_result()` can potentially propagate a task-internal exception (or raise a funcx/globus/network-related exception). And when exceptions occur while getting tasks, the task IDs can also be lost depending on how people have written their code.

So my request would be:
1. Side-effect-free way of querying task status
2. better documentation that the existing operations are NOT side-effect free (note for people used to building distributed systems, this can be quite surprising!)

Start by tracing the existing get_result() behavior and how task status and task IDs are retrieved. Define what a side-effect-free status query must return when task-internal, funcx, Globus, or network exceptions occur. Done means the new query exists and the documentation clearly warns that existing operations are not side-effect free.

Written by the indexing model from the issue text.