Fix CI wrongly validating and testing unrelated integrations on stacked PRs

[vitkyrka]

What does this PR do?

Fixes CI running the full integration test matrix (hundreds of jobs) instead of just the integration a PR actually touches, when the PR's base branch is itself another unmerged (stacked) branch — as seen on #24487, whose base is vwhitchurch/docker-run-env-vars-e2e.

Confirmed root cause: compute-matrix.yml's "Fetch merge base" step computes the test matrix's diff base from the pull_request event's cached base.sha. That value is a snapshot taken the last time this PR's own head was synchronized — it does not update when the base branch itself receives new commits. For a PR stacked on a moving/WIP branch, that cached base.sha can go stale relative to the base branch's real tip, producing a much wider diff than the PR's own changes and triggering the full integration test matrix.

I verified this directly against #24487's failing run (29333935437): the "Fetch merge base" step used base.sha = 283cdf2 (from 2026-07-10), four days older than the base branch's actual tip at run time, producing a huge diff (.github, datadog_checks_dev, gearmand, pulsar, ray, temporal) and running the full matrix. I then pushed a fresh commit to #24487's head (a diagnostic commit, reverted immediately after) — that refreshed the event's base.sha to the base branch's current tip, and the very next run correctly scoped the matrix to just gearmand, with no code change needed. That confirms the staleness is the mechanism; this PR removes the dependency on that lucky-timing refresh by resolving the base branch live every run.

Fixed by diffing against the base branch name (github.event.pull_request.base.ref, percent-encoded) instead of base.sha, so GitHub always resolves its current tip rather than a cached value.

Also included, as an independent correctness fix (not confirmed to have caused the specific validate failure on #24487 — see note below): files_changed()/content_changed() in datadog_checks_dev/datadog_checks/dev/tooling/git.py hardcoded a diff against master/origin/master instead of the actual PR base ref, unlike get_base_ref() (already defined in the same file) which correctly resolves it via GITHUB_BASE_REF. This makes every integration touched by an unmerged base branch look "changed" for ddev validate all changed, which is real and reproducible (confirmed via diagnostics on #24487), but in practice it didn't cause a validation failure there — the per-file license-header content comparison already uses the correct base ref and the affected files were identical, so it passed. Still, running full validations against integrations the PR never touched is wasteful and is a latent correctness gap, so it's fixed the same way (get_base_ref()).

Motivation

PR #24487 intermittently ran the full integration test matrix instead of just gearmand. Root-caused via a scratch worktree plus temporary, reverted diagnostic commits on #24487 (with the author's OK) to confirm the mechanism against live CI rather than guessing. Note: the license-headers failure originally observed on #24487 does not appear to reproduce from the base-ref-selection issue — re-running validate on the same commit passed, so that specific failure looks like a one-off flake rather than something this PR's files_changed() change addresses. This PR only changes diffing/base-ref logic; no code under datadog_checks/ is affected.

Review checklist (to be filled by reviewers)

• Feature or bugfix MUST have appropriate tests (unit, integration, e2e)
• Add qa/required if this PR needs QA validation, or qa/skip-qa if it does not. Exactly one of the two is required.
• If you need to backport this PR to another branch, you can add the backport/<branch-name> label to the PR and it will automatically open a backport PR once this one is merged

[dd-octo-sts]

Validation Report

All 21 validations passed.

Show details

Validation	Description	Status
agent-reqs	Verify check versions match the Agent requirements file	✅
ci	Validate CI configuration and code coverage settings	✅
codeowners	Validate every integration has a CODEOWNERS entry	✅
config	Validate default configuration files against spec.yaml	✅
dep	Verify dependency pins are consistent and Agent-compatible	✅
http	Validate integrations use the HTTP wrapper correctly	✅
imports	Validate check imports do not use deprecated modules	✅
integration-style	Validate check code style conventions	✅
jmx-metrics	Validate JMX metrics definition files and config	✅
labeler	Validate PR labeler config matches integration directories	✅
legacy-signature	Validate no integration uses the legacy Agent check signature	✅
license-headers	Validate Python files have proper license headers	✅
licenses	Validate third-party license attribution list	✅
metadata	Validate metadata.csv metric definitions	✅
models	Validate configuration data models match spec.yaml	✅
openmetrics	Validate OpenMetrics integrations disable the metric limit	✅
package	Validate Python package metadata and naming	✅
qa-label	Validate the pull request declares whether it needs QA for the next Agent release	✅
readmes	Validate README files have required sections	✅
saved-views	Validate saved view JSON file structure and fields	✅
version	Validate version consistency between package and changelog	✅

View full run

[datadog-datadog-us1-prod]

 

🎉 All green!

🧪 All tests passed
❄️ No new flaky tests detected

🎯 Code Coverage (details)
• Patch Coverage: 97.44%
• Overall Coverage: 67.50% (-20.78%)

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: cac2ed1 | Docs | Datadog PR Page | Give us feedback!}