Enhance Alertmanager proxy manifests, Cortex deployment values

[rastislavs]

• enlarge timeout for requests to Alertmanager authorization server, return 503 error when it is unavailable
• use 2 Replicas for Alertmanager, so that we don't produce errors in KKP controllers when its pod is restarted
• enable Ruler, so that we can actually have some alerts in the Alertmanager
• use default replica counts for each component in Cortex write path
• use statefulset for ingesters with the hope for better stability on restarts
• remove unused code/scripts