While investigating T301511 , we discovered that cross-cluster replication settings are incorrect for the Elasticsearch eqiad omega cluster instance (and probably more, will investigate and update this ticket with full scope). Specifically, omega is using itself as a remote cluster seed, but does not have the chi instance configured.

We suspect that this is causing certain searches to err, as recorded in logstash .

Creating this ticket to fix the settings in our running cluster instances, and to get them into puppet for consistency.