Some notes on upgrading the Elasticsearch Universe

In chronological order:

  • Upgrade Monitoring server
  • Perform Kibana snapshot
  • Start backup Kibana instance with the new target version (blue) - the failover for the Load Balancer
  • Upgrade master nodes first
  • Upgrade client nodes
  • Disable shard allocation for all for rolling restarts
  • Upgrade all data nodes one by one
  • Stop primary Kibana instance (green)
  • Upgrade primary Kibana instance to new target version