[Test Plan] [Metrics UI] Infrastructure Anomaly alert type

[Zacqary]

Summary

This test plan covers the new Infrastructure Anomaly alert type (#89244)

See the How to test section of the linked PR for instructions on how to get anomaly data locally.

Scenarios

UI

• An anomaly alert can be created using the Alerts dropdown when on the Inventory page
  • When creating an anomaly alert from the Inventory page, the alert flyout is pre-filled with the corresponding node type if able (anomalies can only be detected on Hosts or Kubernetes pods; other node types should default to Hosts)
• An anomaly alert can be created using the Alerts dropdown when on the Metrics Explorer page
• An anomaly alert can be created from the Alerts Management page
• An anomaly alert can be edited from the Alerts Management page
• The Alert Preview function returns expected results
• The text box for the influencer filter autocompletes one term successfully, and doesn't do anything weird like prompting you to add any AND statements or additional field names

Alert execution

• The alert executes as expected when a new anomaly occurs
• The influencer filter works with an exact match name, e.g. kubernetes-pod-id-abc123
• The influencer filter works with a wildcard, e.g. kubernetes-pod-id-*

[elasticmachine]

Pinging @elastic/logs-metrics-ui (Team:logs-metrics-ui)

[neptunian]

Testing on BC2 pointed at release-oblt, Mac OS Chrome

• When trying to create an infrastructure alert for Kubernetes Pods I get this error if i don't have k8s jobs enabled. It should not mention these jobs but rather the Kubernetes job. If i disable the host job and try to preview the alert the error messaging is accurate.

• When trying to create an infrastructure anomaly alert from the Alerts Management page I get this error. It doesn't happen for other alert types

• If I go to Alerts Management and try to edit an alert I get this error:

[Zacqary]

Looks like the Spaces plugin, which is necessary to create and manage anomaly alerts, isn't enabled in triggers_actions_ui, which wraps Alert Management.

Adding to that difficulty, it doesn't look like we can actually import some kind of context fetcher directly from the spaces plugin to get the current active space. infra loads Spaces services into our useKibana wrapper.

Seems a little complicated to solve this, but I'm looking into it. Definitely need to get this fixed before pushing this alert type out.

[neptunian]

Not sure if your fix made it into BC3 but still get the same results. Will try again on BC4.

[simianhacker]

We decided to pull this feature