Deploying the userdetails application currently requires a developer and someone with admin access in the production AWS account to coordinate multiple manual steps.
One of the 2 userdetails servers must be manually removed all the ALB target groups
Once it's been verified that all connections have drained the application must then be manually deployed from the developers laptop using Ansible to just the server that has been removed from the target groups
When that is complete the server must be manually added back into all the target groups, then it must be verifies that it's in healthy state
Then the whole process is repeated for the 2nd server
This deployment should be automated so there are no manual steps and no administration privileges are required. It's entirely possible to do this with the current server architecture but it's going to save duplicating work to move this straight to containers.
The new process must include:
Single button deploy
Single button rollback
No downtime during deploy
No dependency on the developers local machine or environment, all builds and deploys must be done on version controlled cloud platforms
All code and configuration including hardware setup are checked in and reviewed with the code
Hardware must be auto scaling, high availability and fault tolerant
Automated deployment notifications with including info on environment, deployer, PR
Deploying the userdetails application currently requires a developer and someone with admin access in the production AWS account to coordinate multiple manual steps.
This deployment should be automated so there are no manual steps and no administration privileges are required. It's entirely possible to do this with the current server architecture but it's going to save duplicating work to move this straight to containers.
The new process must include: