A very large request spike to our data replication layer (ElasticSearch) caused the service to start rejecting some incoming requests. This led to data appearing as “not visible” in certain areas of the application for a small period of time, namely register tables and dashboards.