Failed reindexes are fairly common in our Elastic environment. While they're not cause for alarm, they do cause our clusters to dip into red status during routine maintenance operations, such as restarts or reboots.
Our rolling-operation cookbook stops when it detects the cluster is red (which is good!) but it requires manual intervention to clean up the failed indices. The cirrussearch extension repo already has a Python script that detects the failed duplicate indices, so let's make use of this into the rolling-operation cookbook.
AC:
- Rolling operation cookbook detects failed duplicate indices before maintenance operation and prompts user to delete them.