Tracking of migration related issues

[tylersmalley]

• Kibana server is not ready yet #25806 - With a single Kibana instance and after deleting all .kibana_* indices, the migration is reporting that another migration is in progress after creating .kibana_1
• 157050 resource_not_found_exception exception on tasks API with ES basic license

Resolved:

• [migration v6.5] recreate .kibana index  #25796 Need to set "cluster.routing.allocation.enable": null

Possible improvements:

• Messaging to encourage users not to interrupt migration process
• During the initial migration both .kibana_1 (reindex) and .kibana_2 (migration) are created. When re-starting Kibana to retry, the messaging only gives a single index to delete. We should either give both index names, or more preferably only create a single index at a time.
• Increase verbosity of log output for use in debugging
• Add ability to cancel current migration and capture SIGINT.
• Index status/failure into target index.

[elasticmachine]

Pinging @elastic/kibana-operations

[ppf2]

Adding #26049: Kibana not gracefully handling ES cluster health status change (node restarts) during migration

[tylersmalley]

Closing this out-dated meta-issue. Most issues have been resolved and there is a large effort underway to change how migrations work with a priority on minimizing their production impact.