Github user acaciocenteno commented on the pull request:
https://github.com/apache/trafficserver/pull/106#issuecomment-54376276
Thank you for your comments, as I'm just starting to hack with ATS code,
there're certainly many aspects I'll have to learn yet.
But, all in all, do you guys think it's worth adding the feature? If so,
should I go on using UUIDs or just plain sm_ids? I still see some advantages on
the UUID side. For one, if we have a pool of servers and collect their logs to
a loghost, using sm_ids won't be unique, and some further analysis would have
to be done to filter the specific request. Sure, we could add more info to the
sm_id, say, the process' pid, but then why not just use the UUID, which is
guaranteed to be unique?
Also, if I should go on, should I remove the support for adding the id to
the proxy's request and to client's response? Why having that option would not
be good? I mean, it's totally optional and would probabily be used only on very
specific debugging scenarios. What about using Grando's idea of adding the id
to the already existing X-Debug header?
If you guys think it would be worth to go on, I sure can make the necessary
modifications. Thank you.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
