v2.2.3 (hotfix for autoscaling groups)

• #327 added support for autoscaling groups (no changes to Spark code)
• deployment procedure changed, see updated docs at https://github.com/Midburn/Spark/blob/master/docs/development/releases-and-deployment.md#deploying-a-published-release

deployment

• the new load balancer the-real-spark-production is running and working properly
• change the DNS to point to the new load balancer
• during DNS propogation the old instance will still work - so no down-time is expected
• after 48 hours can take the old instance down