Hi Petteri,

That's definitely a bug; Usergrid should never return HTTP 500 internal
server error unless Cassandra or ElasticSearch is inaccessible. Is there a
stack trace that goes along with that error?

Dave


On Tue, Jun 14, 2016 at 10:38 AM Petteri Sulonen <
[email protected]> wrote:

> Hi, everybody --
>
> I'm evaluating Usergrid for our cloud services, and am starting to come
> to grips with it. I have a question about error handling.
>
> I've noticed that the errors Usergrid returns seem somewhat
> inconsistent. For example, when attempting to GET an entity that doesn't
> exist directly from the collection (omitting org and app from path for
> clarity):
>
> GET /things/ferrari
>
> will return an error "entity_not_found", but attempting to GET the same
> entity through a connection where I "own" the entity:
>
> GET /users/me/owns/things/ferrari
>
> will return an "uncaught" error, with error_description "Internal Server
> Error."
>
> Is this intentional? I was expecting to get an "entity_not_found" error
> in both cases, as I think this would be a pretty common situation to
> have in a system where entities are created and deleted dynamically.
>
> (If /things/ferrari actually exists and I own it, both queries return it
> as expected, so that's cool.)
>
> If I've understood the design intent behind Usergrid correctly, these
> connections are extremely important because I can map permissions to
> them. Therefore it's not always possible to access objects directly
> under the collection, as the user may only have permission to access
> objects through a connection (e.g. "owns"). I like this scheme a lot,
> it's simple, intuitive, yet powerful.
>
> If I've misunderstood something, I would greatly appreciate being set
> straight.
>
> Thanks in advance,
>
> Petteri
>
>
>

Reply via email to