Session.id semantic convention#215
Conversation
Co-authored-by: jason plumb <[email protected]>
Co-authored-by: jason plumb <[email protected]>
Co-authored-by: Chengzhong Wu <[email protected]>
bryce-b
force-pushed
the
client-sessions
branch
from
ce4be2e to
2574f89
Compare
bryce-b
force-pushed
the
client-sessions
branch
from
d8c013b to
1434ecf
Compare
- removed 'client' prefix - added session duration & start - removed behavior definitions of session
bryce-b
changed the title
|
@open-telemetry/specs-approvers @open-telemetry/specs-semconv-maintainers This is ready for your review! |
|
We have 3 approvals on this and it's been open for months. |
There was a problem hiding this comment.
Usually in conventions, we also describe things like span names, maybe a structure of a trace (in this case modeling a session). Is this something the group discussed and want to work on? Or the idea is to just define the attribute for now?
I'm missing a bit more of context on why this change is needed. Is there any issue/meeting notes/discussion documented that could help understanding the need for this attribute?
Also please add a changelog entry.
|
@joaopgrassi respectfully, this has been discussed at length in the client SIG. Sessions don't really have a data shape themselves (or at least not yet in our mental model), it's more like a sequence of telemetry all grouped together, as identified by a session identifier. Hence this PR. The idea is to get something started so that we can iterate on it. This is not intended to be comprehensive yet. |
|
@breedx-splk all good, the concept of a session is clear to me. I was more interested in for example an issue where the SIG discussions and agreements/plans are documented. I don't participate in the SIG so having that piece of extra context IMO helps the general semconv approvers. |
|
This has several approvals and has been open for a while with no objections. Merging :) |
13 participants
Discussed this in the client-side sig. Added a basic definition to start discussion.