tree c60bf24765aab6a5f8f9dc31258e73daf7b9de89
parent 70c60feea637143d4a3c14c83f1bfeb988ef7bc4
author Serge Bazanski <q3k@hackerspace.pl> 1602338232 +0200
committer q3k <q3k@hackerspace.pl> 1608242996 +0000

ops/monitoring: scrape apiserver, scheduler, and controller-manager

These get scraped by public IP address, which get retrieved via service
discovery in Prometheus (by using the endpoints role on the
default/kubernetes service).

Also drive-by fix cluster prometheus resources - the default
configuration wants at least 3GB of physical memory.

Change-Id: I1eedb19051f62b40613f69e5f0f736d5958acf42