I left Heroku for OpsWorks and shaved 40% off response times

Re:I left Heroku for OpsWorks and shaved 40% off response times by sync on Friday February 07, 2014 @04:16PM (#7198787)

Seems like Opsworks does a lot more than just the provisioning, no?

"AWS OpsWorks can scale your application using automatic load-based or time-based scaling and maintain the health of your application by detecting failed instances and replacing them. You have full control of deployments and automation of each component."

How would you suggest using Ansible to cover that side of things?

[ Parent | on HN ]

Re:I left Heroku for OpsWorks and shaved 40% off response times by wc- on Friday February 07, 2014 @04:29PM (#7198861)

Autoscaling groups with ansible-pull + cloud init to configure instances as they are spun up by AWS is what I've had the most success with. So your autoscaling group spins up a new instance using a prebaked AMI, and on first boot cloud-init will run ansible-pull locally to grab the latest configs and start up services as needed. Alternatively you can just use ansible to bake a new AMI as needed and skip the ansible-pull + cloud-init part altogether.

[ Parent | on HN ]

Re:I left Heroku for OpsWorks and shaved 40% off response times by imperialWicket on Friday February 07, 2014 @05:04PM (#7199068)

This just introduces additional tools/products to the stack, but you can use Ansible with the Netflix OSS stack, such that Ansible provisions your AMIs and Asgard manages auto-scaling and deployment (with very fine-grained control).

[ Parent | on HN ]