Define case sensitivity of map<string, any> keys in Logs Data Model #3889
Conversation
…standard attribute spec
pellared
changed the title
|
This was previously an undefined behavior, the PR defines the behavior now and I think such changes should be allowed for Stable documents. @open-telemetry/specs-approvers any objections to this approach to stability? |
|
Side note: Actually some logging libraries allow duplicate keys (https://go.dev/play/p/fzIswL0Le7j). OTLP (and JSON) allows also passing such data. EDIT: I will create a separate issue (or PR for it). |
Still good to merge this PR, given this detail, for the March's release? |
@carlosalberto It is good to merge (even if we decide to remove it later). EDIT: I created #3931 |
|
@pellared Oops, lint is failing. |
The failure is not caused by the changes from this PR. |
arminru
added
spec:logs
| keys in the map are unique (duplicate keys are not allowed). The representation | ||
| of the map is language-dependent. | ||
| Value of type `map<string, any>` is a map of string keys to `any` values. | ||
| Case sensitivity of keys MUST be preserved. |
There was a problem hiding this comment.
I'm not sure if the wording here is right. "Case sensitivity MUST be preserved" logically means "if something is case sensitive, it should stay case sensitive; if something is case insensitive, it should stay case insensitive". Doesn't it? 🤔
Perhaps a better wording would be:
The keys in the map are case-sensitive and MUST be unique (duplicate keys are not allowed).
|
Changing to draft per #3938. |
|
This PR was marked stale due to lack of activity. It will be closed in 7 days. |
Per #3853 (comment)
Changes
Add clarifications to
map<string, any>taken from https://github.com/open-telemetry/opentelemetry-specification/tree/main/specification/common and make the description more normative.The spec is stable, but I think that it is rather a clarification or/and straightforward standardization.