[Chargeback] 0.4.1 follow-up: fix empty Usage dashboard (sync @timestamp)

[JohannesMahne]

Summary

• Follow-up to [Chargeback] 0.4.1 bulk: ds_type/ds_namespace panels, event.ingested, docs #20479: billing_cluster_cost and billing_realized_pool now sync on @timestamp instead of event.ingested.
• Fixes empty Usage & Cost Allocation dashboard when On-Prem Billing (or seeded) sources do not map event.ingested, which left billing_realized_pool_lookup at 0 docs.

Test plan

• E2E on stack 9.2.x with billing_realized_pool_lookup count > 0
• Usage ES|QL smoke (STATS … BY ds_namespace) returns rows
• Replaces/supersedes push to feature/chargeback-0.4.1-bulk (branch protection blocked further pushes)

Made with Cursor

[github-actions]

✅ Elastic Docs Style Checker (Vale)

No issues found on modified lines!

---

The Vale linter checks documentation changes against the Elastic Docs style guide. To use Vale locally or report issues, refer to Elastic style guide for Vale.

[infra-vault-gh-plugin-prod]

💔 Build Failed

• Buildkite Build
• Commit: 1dc11f9

Failed CI Steps

• :bash: Dev scripts unit tests

History

[mergify]

Tick the box to add this pull request to the merge queue (same as @mergifyio queue).

• Queue this pull request

[github-actions]

TL;DR

Buildkite failed in :bash: Dev scripts unit tests because the release-commit test hardcodes 10-character short SHAs, but get_release_commit.sh returned 11-character abbreviations in CI. This is a test harness brittleness issue, not a failure in the chargeback package changes in this PR.

Remediation

• Rebase this PR onto the latest main (or cherry-pick the recent dev-scripts test refactor) so the brittle hardcoded SHA assertions are replaced with dynamic expectations from a dummy repo.
• Re-run Buildkite and confirm :bash: Dev scripts unit tests passes.

Investigation details

Root Cause

At the merge-base tested by Buildkite, the dev-scripts test runner contains hardcoded expected abbreviated SHAs in .buildkite/scripts/run_dev_scripts_tests.sh:

• unquoted version assertion expects 43bb655db0
• quoted version assertion expects 8b024204a8
• beta-promotion assertion expects fd04de398f

dev/scripts/get_release_commit.sh emits the first token from git log --oneline (awk '{print $1}'), so SHA abbreviation width can vary by environment/config/repository state. In this CI run it returned one extra character for those cases, causing deterministic assertion mismatches.

This PR itself modifies only packages/chargeback/** files (no dev/scripts/** or .buildkite/scripts/** changes), so the failing job is not introduced by the PR diff.

Evidence

• Build: https://buildkite.com/elastic/integrations/builds/47056
• Job/step: :bash: Dev scripts unit tests
• Key log excerpt:
  • FAIL: finds commit for package at packages/<p>/ (unquoted version) — expected '43bb655db0', got '43bb655db05'
  • FAIL: finds commit for package at packages/<p>/ (quoted version) — expected '8b024204a8', got '8b024204a84'
  • FAIL: finds commit for version promoted from beta — expected 'fd04de398f', got 'fd04de398f1'

Verification

• Not run (read-only detective workflow).

Follow-up

If rebasing is not possible immediately, an alternative is to patch the test expectations to avoid fixed-width SHA literals (e.g., derive expected values dynamically as done in the newer dummy-repo test approach).

---

What is this? | From workflow: PR Buildkite Detective

Give us feedback! React with 🚀 if perfect, 👍 if helpful, 👎 if not.

[JohannesMahne]

UX follow-up: pies removed in favor of Workload Breakdown columns — see #20554 (includes this branch's sync fix).

[JohannesMahne]

Superseded by #20554 (full 0.4.1 + sync fix + Workload Breakdown table columns).