Improve Multi-Region Active-Active profile

[falko]

• switch to internal load balancer
• clean up python script
• rename zone to region in python script
• show automation and pure kubectl/helm commands side by side in tables
• JSON logging
• fix zbctl
• check that cluster in expected state, e.g. using jq on zbctl --insecure status --output json
• fix make target to pause exporters, which needs curl
• make bucket optional or move to other profile
• Add decision table
• remove image: latest from camunda-values.yaml
• set Helm Chart version explicitly
• Switch back to official helm chart in README and region Makefiles
• generate the right number of initial contact points in Python, e.g. by using clusterSize and regions as input
• Maybe add workaround for [ISSUE] Broken Zeebe start script for multi-region failBack mode camunda/camunda-platform-helm#1247
• automate creation of firewall rules
• automate cutting the connections by
  • disabling the firewall rule
  • disabling DNS Chaining
  • if needed, restarting the Zeebe pods
• "disable" Elasticsearch exporter for the failed region by routing exporters to a throw-away ES instance in the failover namespace
  • in surviving brokers
  • in failOver brokers
  • in failBack brokers
• enable Elasticsearch exporter for restored region, i.e. throw away the throw-away ES instance and switch back to the original exporter configuration
• make ES exporter names unique to prevent accidental reuse of remaining sequence data in case of a very fast recovery
• fix Multi-Region: Conflicting exporter names result in lost data #82
• split configuration into common and per-region part
• asymmetric configuration for webapps
  • Run Operate & Tasklist in importer-only mode
  • and/or disable ingress
• Add metrics to k8s
• use --context on all kubectl calls
  • explicitly switch kubectl context before running helm