Define Disaster Recovery plan

[briri]

Most of the infrastructure is built using CF.

We need to identify which resources should have a copy stored elsewhere for DR purposes.

• DynamoTable(s)
• S3 bucket that contains DMP narrative documents
• Cognito User Pool?
• Others?

We also need to document the process of restoring the system in another AWS region.

[briri]

Also include certain SSM credentials

[briri]

What happens with the general AWS Account Id and the KMS key used by the account? What are the implications if we need to recovery to a different account?

[briri]

Setup PointInTimeRecovery on the DynamoDB Table.

• IAS is going to investigate the backup/DR options for us.
• Other option is to make the table a 'Global' table and then create a replica that gets synced

S3 bucket

• Bucket also contains the landing page static website code and the swagger site
  • The landing page files should be ok. This will eventually be moved into the new system so will leave this S3 bucket
  • Remove the Swagger site from the stg/prd env entirely
  • Create a separate bucket (and CF distro? need to see if it can have multiple targets) on dev to host the Swagger site
  • Create a ticket for IAS for the prd bucket

[briri]

IAS indicated that PointInTimeRecovery for Dynamo is good for now. They will do some research and discuss with Kevin to see if there is a more appropriate option.

Created a ticket for IAS to setup replication/backup of the 2 CloudFront Distro buckets that store the DMP narrative PDF documents https://github.com/CDLUC3/uc3sys/issues/75

[briri]

RDS database has been setup to mimic the pattern IAS uses in the main account. Have verified that it is successfully taking daily snapshots

[briri]

all of this work is complete