jonathanhle/Devops Definition of Done Thoughts

## Devops Definition of Done Thoughts
•  Releases Managed
    •	You have to have some type of deployable software object before you can do anything.
    •	The object here is the AMI, container, deb, zip, something
•  Deployments are Automated
    •	Now that you have a deployable software object, you have to automate the deployment.
•  APIs for Observability
    •	Both before and after the object is deployed, you're going to have spend some time instrumenting/researching what characteristics need to be exposed for monitoring at runtime.
    •	What should be measured so we can do whatever is needed to ensure the systems keep working optimally
•  Monitoring & Alerting
    •	Using the APIs, logs, whatever from above, implement monitoring and alerting on KPIs
•  Self-Healing
    •	Using whatever you wrote for monitoring to detect when something is bad and fix yourself
•  Highly Available
    •	The system should be able to deal with faults, so it is able to provide a desired level of service.
    •	Active/Passive, Multi Master, etc.
•  Infrastructure is automated
    •	If the code doesn't live in git, the process won't be repeatable; which makes it prone to human error
•  Secure
    •	Store secrets in a smart way
    •	If we can't keep up with security updates, will that put us at risk
•  Long term Support Process
    •	All systems need care and feeding sometimes
    •	How do we make this as painless as possible
	• Releases Managed
	• You have to have some type of deployable software object before you can do anything.
	• The object here is the AMI, container, deb, zip, something
	• Deployments are Automated
	• Now that you have a deployable software object, you have to automate the deployment.
	• APIs for Observability
	• Both before and after the object is deployed, you're going to have spend some time instrumenting/researching what characteristics need to be exposed for monitoring at runtime.
	• What should be measured so we can do whatever is needed to ensure the systems keep working optimally
	• Monitoring & Alerting
	• Using the APIs, logs, whatever from above, implement monitoring and alerting on KPIs
	• Self-Healing
	• Using whatever you wrote for monitoring to detect when something is bad and fix yourself
	• Highly Available
	• The system should be able to deal with faults, so it is able to provide a desired level of service.
	• Active/Passive, Multi Master, etc.
	• Infrastructure is automated
	• If the code doesn't live in git, the process won't be repeatable; which makes it prone to human error
	• Secure
	• Store secrets in a smart way
	• If we can't keep up with security updates, will that put us at risk
	• Long term Support Process
	• All systems need care and feeding sometimes
	• How do we make this as painless as possible