[fix][client] Fix incorrect topic in OTel spans for partitioned producers

[void-ptr974]

Fixes #25897

Motivation

The OTel producer interceptor cached the topic from the first producer it processed.

For partitioned producers, the same interceptor can be used by multiple internal partition producers. For example, if the first message is sent to:

persistent://public/default/my-topic-partition-0

the interceptor caches that topic. If a later message is sent to:

persistent://public/default/my-topic-partition-1

the producer span could still be created as:

send persistent://public/default/my-topic-partition-0

and messaging.destination.name could also be reported as:

persistent://public/default/my-topic-partition-0

This makes OTel producer spans report an incorrect topic for messages sent to other partitions.

Modifications

Use the current producer topic when creating each OTel producer span instead of caching the first topic.

Verifying this change

• Make sure that the change passes the CI checks.

This change added tests and can be verified as follows:

• Added an integration test that sends messages with a partitioned producer and verifies producer spans use the correct partition topics.
• Ran ./gradlew :pulsar-client-original:compileJava
• Ran ./gradlew :pulsar-broker:test --tests org.apache.pulsar.broker.service.OpenTelemetryTracingIntegrationTest.testPartitionedProducerSpanUsesCurrentProducerTopic

Does this pull request potentially affect one of the following parts:

• Dependencies (add or upgrade a dependency)
• The public API
• The schema
• The default values of configurations
• The threading model
• The binary protocol
• The REST endpoints
• The admin CLI options
• The metrics
• Anything that affects deployment

[void-ptr974]

@Technoboy- Could you please take a look at this small OpenTelemetry tracing fix when you have time? Thanks!

[Technoboy-]

@Technoboy- Could you please take a look at this small OpenTelemetry tracing fix when you have time? Thanks!

could you fix the tests

[Technoboy-]

@Technoboy- Could you please take a look at this small OpenTelemetry tracing fix when you have time? Thanks!

you can merge this after the test fixed

[void-ptr974]

The previous CI failure was from the unrelated ReplicatorTest.testResumptionAfterBacklogRelaxed, which was fixed by #26002. I have merged the latest master; both the OTel regression test and the previously failing Replicator test pass locally. Please approve the new CI runs when convenient. Thanks!

[lhotari]

depends on #25585 and #25749 while cherry-picking to branch-4.2. Performing the cherry-picking and backporting now.

[lhotari]

branch-4.0 doesn't contain PIP-446, so dropping branch-4.0 as the backporting target.