No-one likes being woken up with a PagerDuty alert at 4am for a problem which could resolve itself with no human intervention, but for looking at it a bit differently. When moving existing applications into the cloud, or designing new ones targeting the cloud, it's important to consider and take advantage of the capabilities of the target platform, while not locking yourself in to a particular vendor.
In this talk we'll look at a variety of different types of software and look at real world deployment patterns for running them in a self-assembling, self-healing manner in the Amazon cloud, and about how we might be able to replicate some of these capabilities on other cloud platforms.