The following shows how B2B consultants created a fully automated AWS disaster standby system for a company with a hybrid on-premise / AWS Cloud infrastructure.
Please see the Figure 1 below for an overview of how it works:
In summary it operates as follows:
- AWS Route53 based Health Checks monitor both the Live Service and the Disaster Standby Service
- If the Health Checks determine that the Live Service has gone down, they raise an AWS CloudWatch Alarm which in turn both automatically redirects the Live Service from the Primary DNS records to the Secondary DNS records (the Secondary DNS records will initally be pointing towards the Maintenance Service), and also using AWS Lambda starts the Disaster Standby Service
- A further AWS Health Check monitors the Disaster Standby Service and when it detects that the service is up and ready it raises an AWS CloudWatch Alarm which in turn using AWS Lambda changes the IP addresses against the Secondary DNS records from the Maintenance Service to the Disaster Standby Service
- The Live Service Health Check service will continue to monitor the Live Service. When it detects the Live Service is back up and running it redirects the Live Service back from the Secondary DNS records to the Primary DNS records
- Finally the AWS Health Check that monitors the Disaster Standby Service will detects that the service has been shut down (manually) and using AWS Lambda will change the IP addresses against the Secondary DNS records back to the Maintenance Service
NOTE: The Database Service is based on a Mirrored Microsoft SQL Server Service and is on separate geographically separate servers. Both the Live and Disaster Standby Web Service will point at the whichever database instance is currently live.