On a periodic basis (via a script called from cron), we are pulling both the Impala text and thrift profiles from the CM. It works fine mostly.
Following this:
https://cloudera.github.io/cm_api/apidocs/v13/path__clusters_-clusterName-_services_-serviceName-_im...
During very busy periods, pulling profiles sometimes returns “Unknown profile” errors. We are assuming that the profiles are rolling off.
Is it possible to a) extend the retention period, and/or b) increase the number saved.