Some questions in regards to AWS integration and druid web services

Hi there!

I've been trying to ask this on IRC channel, but it seems there aren't
much people alive...

Couple of questions, if someone can help me:
1. Is there any way to set up URI prefix for every Web service like
broker, coordinator, overlord, etc? So for e.g. to open overlord
console I should enter: http:/some_host:port/overlord/console.html
2. Is there any way to eliminate necessity to define creds in task
payload and use credentials those set in druid's configs?
3. Does Druid supports AWS roles and AWS STS service to obtain temporary creds?

Also one thing not really sounds valid to me:
Druid assumed to offload old segments to deep storage (S3) and well as
sync them from there if data missing, so druid creds should have RW
access to bucket that hosts base and archive and logs. However, when
dispatching a task to ingest some data, you provide credentials to
access that specific location, but then, when index is build, druid
tries to upload index.zip to its own bucket (deep storage) with
credentials provided in task - that's looks weird a bit (or it's a
bug?)

Version in use 0.9.1.1

Any advices or ideas are welcome!
Thanks in advance.