Fix flaky Data Streams assertions in HTTP tests

[AlexeyKuznetsov-DD]

What Does This Do

Adds waitForGroup to RecordingDatastreamsPayloadWriter and uses it in the shared HTTP tests to wait for the expected root StatsGroup before calling verifyAll.

Note: Test changed a lot of places, but they all identical.

Motivation

These tests were flaky on GitLab CI because Data Streams groups are reported asynchronously. waitForGroups(1) only guaranteed that any group existed, so the immediate groups.find { it.parentHash == 0 } could return null before the expected root group arrived. A Develocity retry could pass with a more favorable scheduling order.

waitForGroup waits for the matching group with a bounded timeout, removing that race and producing a clearer failure if the group never arrives.

Additional Notes

Before the fix tests failed on GitLab Ci:

org.spockframework.runtime.SpockAssertionError: Target of 'verifyAll' block must not be null

Target of 'verifyAll' block must not be null
	at spock.lang.Specification.verifyAll(Specification.java:286)
	at datadog.trace.agent.test.base.HttpServerTest.test success with #count requests(HttpServerTest.groovy:685)

Verification:

• Added targeted test: RecordingDatastreamsPayloadWriterTest.
• Manually tested on GitLab CI after the fix - looks good (1 pass).

[datadog-datadog-prod-us1-2]

More details

PR eliminates test flakiness by adding waitForGroup() to properly wait for asynchronous Data Streams group arrival before assertions. Implementation correctly synchronizes access, respects timeouts, and returns matching groups. All 47 test migrations are consistent and safe. No behavioral regressions detected.

Was this helpful? React 👍 or 👎

_{📊 Validated against 5 scenarios · Open Bits AI session}

_{🤖 Datadog Autotest · Commit bc39012 · What is Autotest? · Any feedback? Reach out in #autotest}