[UA] Upgrade assistant migration meta data can become stale

[alisonelizabeth]

Fixes #46146

This PR checks for any indices that have a deprecation warning and also a reindex operation of "completed", and deletes the reindex operation if so.

How to test

1. Modify L46 of es_migration_apis.ts with the following mock data:

    indices: indices.concat([
      {
        level: 'critical',
        message: 'test failed fix',
        url: 'test/url',
        details: 'test details',
        index: 'test',
        reindex: true,
        needsDefaultFields: true,
      },
    ]),

2. Create the test index

PUT test/_doc/1
{"foo":"bar"}

3. Create a snapshot of the test index.
   You can use the Snapshot Restore UI to do this or Console using ES APIs. From the UI:

   • Create a repository
   • Create a policy with the test index selected
   • Execute the policy

4. Go to Upgrade Assistant and reindex the test index

5. Delete the reindexed index

6. Restore the snapshot (via UI or Console)

7. Navigate back to Upgrade Assistant. Verify the UI indicates that it needs to be reindexed again.

Release note

A fix was made to the Upgrade Assistant where the reindexing of an index was incorrectly marked as "Done", when it still required to be reindexed. This scenario could occur if a user reindexed an index, deleted it, then restored a snapshot of the index of an older version.

[elasticmachine]

Pinging @elastic/es-ui (Team:Elasticsearch UI)

[jloleysens]

@alisonelizabeth Great work! Thanks for fixing this bug.

I tested locally and this contribution fixes the stale metadata bug!

[jloleysens]

@elasticmachine merge upstream

[kibanamachine]

💚 Build Succeeded

• continuous-integration/kibana-ci/pull-request
• Commit: 23c61be

History

• 💔 Build #35406 failed 0d530d5
• 💚 Build #35337 succeeded 828ab9a
• 💚 Build #35022 succeeded 4c80938

To update your PR or re-run it, just comment with:
@elasticmachine merge upstream