Chris Kleinknecht

> Do we provide this in the API package or API+SDK? If we can solve the

Thanks for the PR @guewen. Using the `six` types to solve the problem of exporting `unicode` attribute values looks right to me. In general though this problem is a big...

Which is all to say: the direction looks good, but there may be some unintended consequences.

There's some pending work to support pubsub/task queues/etc. in general, but nothing concrete yet as far as I know. It looks like the datadog integration persists trace data for celery...

This looks like a good solution to me too pending the TODOs in the description.

@marctc thanks for raising the issue, and sorry for the long delayed response here. Can you still reproduce this with recent versions of `opencensus` and `opencensus-ext-django`?

After talking to @bogdandrutu we'll wait on this and #426 until adding support for [resources](https://github.com/census-instrumentation/opencensus-specs/blob/master/resource/Resource.md).

LGTM with a test case that fails on master, passes with this change.

For (1), the problem is that the requests integration is creating two nearly identical spans, one the child of the other?

This PR has grown in scope a bit as I've tried to quarantine stackdriver logic in the stackdriver exporter. See https://github.com/census-instrumentation/opencensus-python/pull/649#discussion_r283612298 for details.