Determine performance hit of getting caller thread stack trace on every command execution

[mattrjacobs]

See #195

[benjchristensen]

I'm pretty certain we don't want to do that. Also, as we move towards 1.4 and mostly non-blocking execution, the call stack of the caller becomes less and less useful to capture when an error occurs (as the caller could be anywhere). This issue already exists in 1.3 and earlier where if queue() or observe() are used, an error being thrown capture the caller thread wherever it is, but since it's not blocking on the HystrixCommand the stack is rather meaningless.

We may just want to eliminate this behavior, or if it is thread, it's just something optional to do but not automatic.

[mattrjacobs]

Not worth examining. Closing.