chore: release API 1.6.0/SDK 1.17.0/Experimental 0.43.0

[pichlermarc]

Release PR for #4137 (needs rebase after the PR merges). We peer-depend on API <1.6.0 in the SDK packages, therefore we need to release everything.

API 1.6.0

🐛 (Bug Fix)

• Revert "feat(api): add attributes argument to recordException API #4071"
  • This feature was an unintentional breaking change introduced with 1.5.0

SDK 1.17.0

🐛 (Bug Fix)

• Revert "feat(api): add attributes argument to recordException API #4071"
  • This feature was an unintentional breaking change introduced with API 1.5.0
  • This PR updates all SDK packages to allow API 1.6.0, where this change has been reverted.

Experimental 0.43.0

🐛 (Bug Fix)

• Revert "feat(api): add attributes argument to recordException API #4071"
  • This feature was an unintentional breaking change introduced with API 1.5.0
  • This PR updates all experimental packages to allow API 1.6.0, where this change has been reverted.

[codecov]

Codecov Report

Merging #4138 (a6b4375) into main (27897d6) will increase coverage by 2.55%.
The diff coverage is n/a.

❗ Current head a6b4375 differs from pull request most recent head 5be7de4. Consider uploading reports for the commit 5be7de4 to get more accurate results

@@            Coverage Diff             @@
##             main    #4138      +/-   ##
==========================================
+ Coverage   89.68%   92.23%   +2.55%     
==========================================
  Files         140      326     +186     
  Lines        3014     9299    +6285     
  Branches      653     1971    +1318     
==========================================
+ Hits         2703     8577    +5874     
- Misses        311      722     +411     

[dyladan]

Should this be API 1.5.1 or 1.6.0? Anyone depending on ^1.5.0 could see a breaking change by this "fix." Not sure how semver handles making breaking changes when the bug is itself just a semver violation? This would simply be introducing another breaking change. the argument could be made that 1.5.0 should instead be deprecated or something? I'm really unsure.

I'm also unsure that I agree with reverting this change. It is not breaking for end users, only implementers. This is why our SDKs specify a maxiumum API version in the peer dependencies (example: https://github.com/open-telemetry/opentelemetry-js/blob/main/packages/opentelemetry-sdk-trace-base/package.json#L93).

[dyladan]

/cc @ksstoneware for visibility on my above comment

[dyladan]

I'm also unsure that I agree with reverting this change. It is not breaking for end users, only implementers. This is why our SDKs specify a maxiumum API version in the peer dependencies (example: https://github.com/open-telemetry/opentelemetry-js/blob/main/packages/opentelemetry-sdk-trace-base/package.json#L93).

After thinking more about it. Let's go ahead and revert this for now. It's clear that the third party SDK implementers (afaik dd is the only one) didn't have the same expectations we did. In the SIG meeting we'll discuss how we want to deal with changes like this moving forward.

[dyladan]

This PR follows the official semver guidelines

https://semver.org/#what-do-i-do-if-i-accidentally-release-a-backward-incompatible-change-as-a-minor-version

[vinicius73]

Hi @pichlermarc cold you release 1.6.0 at npm?
At now we only have 1.5.0

[BrettHoutz]

It has been released now.

[pichlermarc]

@vinicius73 yes - the release process takes some time but it should be up now.