Fix consumption codec chart (stacked onto #19520)

[JohannesMahne]

Summary

Stacked PR to land dashboard fixes onto feature/es-index-mode-codec-compat for #19520.

• Fix Primary ingest by codec over time Lens xAccessor (invalid column ID after cloning from mode panel)
• Replace mode/codec table with codec-over-time chart on consumption dashboard
• Merge latest Index & Shard View export with Kibana 8.10 compatibility fixes

Stacked onto #19520

Test plan

• CI green on this stacked PR
• Merge into feature/es-index-mode-codec-compat
• Verify elasticsearch: add index mode and codec dashboard visualizations #19520 shows the new commit

Made with Cursor

[JohannesMahne]

@elastic/stack-monitoring - sorry about the noise. I am busy with another PR, but the repo blocked me from adding commits to it. I am not sure how to prevent this behaviour. Your advice will be appreciated.

[consulthys]

Totally fine @JohannesMahne let us know when you need some review/approval 👍

[elasticmachine]

💔 Build Failed

• Buildkite Build
• Commit: 8207d28

Failed CI Steps

• Check integrations elasticsearch

History

[JohannesMahne]

@consulthys - if you can review this one, so that it can be part of #19520 - that would be great.

I have a review meeting on #19520 later the week for the change, so it would be great if it can be the complete PR with all the changes. There might have one or 2 small dashboard changes after the review that needs to be added.

[consulthys]

LGT Stack Monitoring

[github-actions]

TL;DR

Check integrations elasticsearch failed, but the prefetched Buildkite log only contains teardown plus exit status 1, not the actual failing command output. I could not identify a code-level root cause from the available CI data; the next action is to rerun the Buildkite step or inspect the full private job log/artifacts.

Remediation

• Rerun Check integrations elasticsearch, or open the full Buildkite job log for the failing step and look above teardown for the first elastic-package install / elastic-package test error.
• If the failure repeats, share the full failing excerpt from before --- [elasticsearch] failed; the currently available log tail is insufficient to trace the failure to source.

Investigation details

Root Cause

Inconclusive from the available logs. The changed PR files are dashboard JSON files under packages/elasticsearch/kibana/dashboard/, and local static validation of those PR dashboard files did not reproduce a failure. The stack-backed asset/import test could not be run in this environment because Docker is unavailable.

Evidence

• Build: https://buildkite.com/elastic/integrations/builds/44479
• Job/step: Check integrations elasticsearch
• Failed command: .buildkite/scripts/test_one_package.sh packages/elasticsearch origin/feature/es-index-mode-codec-compat 8207d28f39de8f5d2498f2ea9d8cad47f8d6db23
• Available log excerpt:

  --- [elasticsearch] failed
  ^^^ +++
  🚨 Error: The command exited with status 1
  ^^^ +++
  user command error: exit status 1
  

Verification

• Ran elastic-package check -v on a temporary package copy with the two PR dashboard files applied: passed.
• Ran elastic-package test static -v --report-format xUnit --report-output file on the same temporary package copy: passed.
• Ran a deterministic Lens reference check over the PR dashboard JSON for missing layerId, xAccessor, splitAccessor, accessors, yConfig.forAccessor, columnOrder, and formula/order references: passed.
• Could not run elastic-package test asset locally because it requires a running Elastic stack/Docker.

Follow-up

If the rerun fails again, the actionable excerpt should be the first error before the teardown block, likely in package installation, asset loading, or stack-backed test execution.

---

What is this? | From workflow: PR Buildkite Detective

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