test(debugger): budget limit for capture expressions

[tylfin]

Motivation

Extends the Live Debugger budget tests to ensure that probes configured with capture expressions respect the 1 snapshot per second limit, the same budget already enforced for full-snapshot log probes.

Changes

• New probe fixture probe_capture_expressions_budgets.json: a line probe on the Budgets method with captureSnapshot: false and a captureExpressions entry on the loops parameter.
• New test Test_Debugger_Probe_Budgets::test_log_line_capture_expression_budgets: fires 150 requests and asserts that only 1-20 snapshots carrying capture expressions are emitted (matching the existing test_log_line_budgets tolerance for the ~1/s budget).
• Marked the test as a known bug for Java (DEBUG-5730) in manifests/java.yml: capture-expression probes currently emit on every invocation (150/150) and are not yet rate limited.

Testing

Ran DEBUGGER_PROBES_SNAPSHOT locally for Java:

• Without the manifest entry, the test correctly fails (got 150 ... should be 1-20), confirming the budget is bypassed.
• With the DEBUG-5730 bug declaration, the test is skipped as expected.

🤖 Generated with Claude Code

[github-actions]

CODEOWNERS have been resolved as:

tests/debugger/utils/probes/probe_capture_expressions_budgets.json      @DataDog/debugger @DataDog/system-tests-core
manifests/java.yml                                                      @DataDog/asm-java @DataDog/apm-java
manifests/php.yml                                                       @DataDog/apm-php @DataDog/asm-php
manifests/python.yml                                                    @DataDog/apm-python @DataDog/asm-python
tests/debugger/test_debugger_probe_budgets.py                           @DataDog/debugger @DataDog/system-tests-core

[datadog-official]

🎉 All green!

🧪 All tests passed
❄️ No new flaky tests detected

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: bfc6cda | Docs | Datadog PR Page | Give us feedback!}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: e4490c0cc6

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Add the new test to the scenario map

This new test is not added to tests/test_the_test/scenarios.json, which is the map loaded by utils/scripts/compute_libraries_and_scenarios.py to translate manifest nodeids and changed test files into scenarios. Because the new manifest entries use this exact nodeid, future PRs that only adjust this test’s manifest status will not select DEBUGGER_PROBES_SNAPSHOT, so the CI impact calculation can silently miss the scenario. Please regenerate/update the scenario map entry for this test.

Useful? React with 👍 / 👎.

[cbeauchesne]

I'm slightly concerned about the stability of this test. From experience, every time we;ve introduced time-based assertions, we've eventually been hit by the fact that we don;t control the underlying machine the test runs on, which introduces a lot of non-determinism.

In the end, we often had to expand the asserted window significantly to reduce flakiness, which defeats the purpose of the assertion.

Let's give this a try, but if it turns out that we get too many failures that cannot be correlated with a real issue, we'll need to rethink this approach.

[tylfin]

AFAIK the lack of stability for the debugger tests was always a missing waiting signal, we should be fairly deterministic as long as we gate on the proper signals

[cbeauchesne]

Approved, with one comment :)