[ILM] Surface linked hidden indices for policy

[alisonelizabeth]

Fixes #87897

An ILM policy associated to a hidden data stream would show "0" for linked indices and allow deletion in the UI, when it should be disabled. This bug does not exist on 7.10, so skipping release note. I think it is related to the introduction of hidden data streams via elastic/elasticsearch#63987.

How to test

This can be tested against the ilm-history-ilm-policy or the watch-history-ilm-policy policies. You can also create your own policy with a hidden data stream as outlined below:

1. Create a hidden data stream, using the following commands in Console:

PUT _index_template/my_template
{
  "template": {},
  "index_patterns": [
    "foo"
  ],
  "data_stream": {
    "hidden": true   
  }
}

POST foo/_doc
{
  "@timestamp": "2020-01-27"
}

2. Create an ILM policy through the UI (only name is required)
3. Go to Index Management, find the backing index for the data stream created in step 1 and open the details panel. Click "Manage" and select "Add lifecycle policy".
4. Go back to ILM UI and verify the policy shows there is a linked index and deletion is disabled.

[elasticmachine]

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

[kibanamachine]

💚 Build Succeeded

• continuous-integration/kibana-ci/pull-request
• Commit: 9deec9d

Metrics [docs]

✅ unchanged

History

• 💚 Build #98126 succeeded 1fd9caf42ed21544dee41fcf231a3c3088dce81d

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

[yuliacech]

Thanks for working on this, @alisonelizabeth ! I tested locally and the fix worked. Great job adding API integration tests too 👍
While I was testing though, I noticed that while the link to Index Management works, the indices list there is not including the correct ILM information for backing indices of a hidden data stream. I believe the same fix is needed here. I'll create an issue for that.

[alisonelizabeth]

While I was testing though, I noticed that while the link to Index Management works, the indices list there is not including the correct ILM information for backing indices of a hidden data stream. I believe the same fix is needed here. I'll create an issue for that.

Great catch @yuliacech! I think you're correct. Thanks for opening up an issue for this.