In a disaster scenario, manual intervention can be used to restore the services on the cloud or on premises, but automation is always best. To automate the failure response, you can use AWS Lambda, by performing a GetHealthCheckStatus request on the health check object of Route 53, which represents your primary web application. Lambda scheduled functions are designed to provide a minimum of one-minute granularity for execution, plus the health check request interval (the standard is 30 seconds, and the fast version is 10 seconds).
The Lambda function can programmatically trigger the following actions:
- Creating the RDS instance
- Creating an EBS Volume
- Creating an EC2 instance with the DR AMI
- Associating the Elastic IP to the running instance
The EC2 instance will perform the following activities:
- Attaching the EBS volume to the instance, via bootstrapping
- Downloading the database backup file from S3 in the instance store, and execute it remotely, using the RDS endpoint
Some of the relevant services are as follows:
- IAM
- Route 53
- Lambda
- EBS
- EC2
- RDS
- CloudWatch
- Direct Connect
- Storage Gateway
- Import Export Snowball