feat(storage): add resource span attributes for ACO ( App Centric Observability ) for rewrite API in async client

[bajajneha27]

No description provided.

[gemini-code-assist]

Code Review

This pull request adds support for span enrichment in RewriteObject within the asynchronous connection tracing implementation. It introduces an overload of MakeTracingAsyncRewriterConnection that accepts an OpenTelemetry span, allowing the span to be created and enriched with destination bucket details before being passed to the tracing connection. A unit test has been added to verify that the span is correctly enriched using cached bucket metadata. I have no feedback to provide as the changes are well-implemented and include appropriate test coverage.

[codecov]

Codecov Report

❌ Patch coverage is 97.43590% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 92.31%. Comparing base (d799c3f) to head (73a7ed6).

Files with missing lines	Patch %	Lines
.../storage/internal/async/connection_tracing_test.cc	96.55%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #16285      +/-   ##
==========================================
- Coverage   92.31%   92.31%   -0.01%     
==========================================
  Files        2223     2223              
  Lines      208086   208121      +35     
==========================================
+ Hits       192086   192117      +31     
- Misses      16000    16004       +4     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[bajajneha27]

/gcbrun

[kalragauri]

There is a limitation in our overall ACO caching design. Because we call EnrichSpan() immediately at span creation across the async client, a background cache miss results in the span never being enriched. For long-running operations, the cache populates milliseconds later, but the span stays unenriched. In the Java SDK, they avoid this by wrapping spans in an AcoSpan decorator that lazily pulls from the cache right at span.end(). Let's track this as a separate issue.

[kalragauri]

There is a pre-existing issue here that we should probably track as a follow-up.

Currently, in AsyncRewriterTracingConnection::Iterate() (and similarly for .Write() and .Read() in the async client), we call span->AddEvent(...) for every single loop iteration.

The risk here is that OpenTelemetry SDKs enforce a limit on the number of events per span. For large object rewrites, .Iterate() can execute hundreds or thousands of times. Once we exceed the limit, OTel silently drops the rest, affecting observability for the tail end of any large data transfer.

Other Storage SDKs have managed to avoid this issue, so let's chat about how C++ can also avoid it.

[kalragauri]

nit: pls fix the PR title + description.

[bajajneha27]

There is a limitation in our overall ACO caching design. Because we call EnrichSpan() immediately at span creation across the async client, a background cache miss results in the span never being enriched. For long-running operations, the cache populates milliseconds later, but the span stays unenriched. In the Java SDK, they avoid this by wrapping spans in an AcoSpan decorator that lazily pulls from the cache right at span.end(). Let's track this as a separate issue.

Ack! I've created issues to fix both the concerns.

[bajajneha27]

/gcbrun